Files
libakstdlib/src/stdlib.c

1343 lines
51 KiB
C
Raw Normal View History

#include <akstdlib.h>
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
#include <ctype.h>
#include <errno.h>
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
#include <limits.h>
#include <stdarg.h>
2026-06-02 17:11:38 -04:00
#include <stdint.h>
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
Wrap the strings, streams and collections the wishlist asked for TODO.md section 3.1's high-priority list and section 3.6's data structures. This is the surface akbasic went without: it makes 10 calls into this library and 116 to raw libc, and 69 of those 116 are strlen, strcmp, strncpy and strstr. Three new translation units, because src/stdlib.c covering four times what it did would stop being readable: src/string.c lengths, bounded copy and concatenation, duplication, comparison including the case-insensitive forms, searching, the reentrant tokenisers, and a status message that knows this library's own statuses as well as errno's. src/stream.c positioning, flushing and buffering, character and line I/O, stream state, freopen/fdopen/tmpfile, formatted input, and the file operations. src/collections.c the list functions that were missing, a head/tail container so append is O(1), a binary search tree, FNV-1a, a fixed-capacity hash map and a growable string buffer. Two conventions run through all of it. The copying functions take the destination size even where the libc function they are named for does not, because strcpy(3) cannot be called safely without it, and truncation is an error that writes nothing rather than a plausible prefix -- this is the idiom akbasic writes out by hand at ten sites. The searching functions treat "not found" as a successful answer of NULL, because absent is an answer and raising on it would make every caller handle a non-error. The hash map is akbasic's src/symtab.c generalised: open-addressed with linear probing over a caller-supplied slot array, keys copied into fixed slots so the map owns them, tombstones on delete so a removal cannot cut a probe chain, and a refusal rather than a resize when full. Tests: 16 binaries, green under the normal and sanitizer builds. The scanf wrappers take the number of conversions the caller expects, since comparing scanf(3)'s return against that by hand is the check everyone eventually forgets. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:32:35 -04:00
#include "aksl_internal.h"
Version the library at 0.1.0 project() now carries VERSION 0.1.0, and is the single place a version number is spelled. It flows into generated version macros, the shared library's VERSION/SOVERSION, the Version: field in akstdlib.pc, and a new akstdlibConfigVersion.cmake. Before this @PROJECT_VERSION@ expanded to nothing, so akstdlib.pc shipped an empty Version: and libakstdlib.so carried no soname at all. 0.x on purpose: TODO.md section 2.1 still records four confirmed defects whose fixes change documented behaviour, so the API is not being promised yet. While the major version is 0 the soname carries MAJOR.MINOR -- 0.1 and 0.2 are different ABIs -- and becomes MAJOR alone at 1.0. The if() in CMakeLists.txt and the #if in tests/test_version.c encode that rule and are tested against each other. include/akstdlib_version.h.in is configured into the build tree as akstdlib_version.h and installed beside akstdlib.h. It defines AKSL_VERSION_MAJOR/MINOR/PATCH/STRING/NUMBER and AKSL_VERSION_SONAME. AKSL_VERSION_NUMBER is computed rather than written as a literal, because a literal 000100 is octal in C and would make 0.1.0 compare as 64; test_version.c asserts it against the runtime components, so a rewrite to a literal fails. Those macros record what a caller was compiled against. aksl_version(), aksl_version_string() and aksl_version_soname() report what actually loaded, and AKSL_VERSION_CHECK() compares the two, raising AKERR_VALUE naming both. It is a macro so that it expands at the caller's site and captures the caller's numbers; the function compares them against the ones baked into the library. Compatibility is "same soname", so patch is ignored -- a caller built against 0.1.0 keeps working against 0.1.7. Normally the soname catches a mismatch at load time and the check never fires. It earns its keep when the soname is bypassed: a 0.2.0 build dropped in under the 0.1 filename loads happily, and only the check notices. write_basic_package_version_file() uses SameMinorVersion to mirror the soname, falling back to ExactVersion below CMake 3.11 where that mode does not exist. The fallback is stricter than the soname rule -- it pins the patch level too -- but never laxer, and wrongly refusing a good pairing beats wrongly accepting a bad one. Coverage of src/stdlib.c rose to 99.1% of lines (217/219), 45.1% of branches and 25/25 functions. That puts branch coverage back over the old 45 gate, but the gate stays at 40: 0.1 points of headroom is not a ratchet. ctest 14/14, ASan+UBSan 14/14, coverage 16/16 at 90/40. Also verified out of tree: SONAME libakstdlib.so.0.1 recorded in consumers, pkg-config --modversion reporting 0.1.0, find_package(akstdlib 0.1) accepted with 0.2 and 1.0 refused, a patch-bumped 0.1.1 loading and passing the check, a 0.2.0 dropped in under the 0.1 filename caught by it, and an embedded add_subdirectory build keeping its own version rather than the parent's. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 22:40:49 -04:00
/*
* Version of the library itself. These read the AKSL_VERSION_* macros as they
* were when *this translation unit* was compiled, which is what makes them the
* loaded library's version rather than the caller's: the caller's copy of the
* same macros comes from whatever akstdlib_version.h it compiled against.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_version(int *major, int *minor, int *patch)
{
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, major, AKERR_NULLPOINTER, "major=%p, minor=%p, patch=%p",
(void *)major, (void *)minor, (void *)patch);
FAIL_ZERO_RETURN(e, minor, AKERR_NULLPOINTER, "major=%p, minor=%p, patch=%p",
(void *)major, (void *)minor, (void *)patch);
FAIL_ZERO_RETURN(e, patch, AKERR_NULLPOINTER, "major=%p, minor=%p, patch=%p",
(void *)major, (void *)minor, (void *)patch);
*major = AKSL_VERSION_MAJOR;
*minor = AKSL_VERSION_MINOR;
*patch = AKSL_VERSION_PATCH;
SUCCEED_RETURN(e);
}
const char *aksl_version_string(void)
{
return AKSL_VERSION_STRING;
}
const char *aksl_version_soname(void)
{
return AKSL_VERSION_SONAME;
}
/*
* Compatibility is "same soname". Pre-1.0 that is MAJOR.MINOR, so both are
* compared and patch is deliberately ignored; from 1.0 the minor comparison
* comes out, in step with the SOVERSION expression in CMakeLists.txt. The
* caller's numbers arrive as arguments because AKSL_VERSION_CHECK() expanded
* them at the caller's site -- see the macro in akstdlib.h.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_version_check(int major, int minor, int patch)
{
PREPARE_ERROR(e);
(void)patch;
FAIL_NONZERO_RETURN(e,
(major != AKSL_VERSION_MAJOR || minor != AKSL_VERSION_MINOR),
AKERR_VALUE,
"compiled against libakstdlib %d.%d.%d, loaded %s (soname %s)",
major, minor, patch,
AKSL_VERSION_STRING, AKSL_VERSION_SONAME);
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/*
* malloc(0) is the awkward case, and it is why *dst is cleared first and errno
* is cleared before the call. The standard lets malloc(0) return either a unique
* pointer or NULL, and a NULL return there is not a failure and need not set
* errno -- so the old code raised an error whose status was whatever errno
* happened to be holding, quite possibly 0. A zero-size request is rejected on
* its own terms instead, since there is nothing useful to hand back either way.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_malloc(size_t size, void **dst)
{
PREPARE_ERROR(e);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
FAIL_ZERO_RETURN(e, dst, AKERR_NULLPOINTER, "dst=%p", (void *)dst);
*dst = NULL;
FAIL_ZERO_RETURN(e, size, AKERR_VALUE, "zero-size allocation");
errno = 0;
*dst = malloc(size);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
FAIL_ZERO_RETURN(e, *dst, AKSL_ERRNO_OR(ENOMEM), "%zu bytes", size);
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
akerr_ErrorContext AKERR_NOIGNORE *aksl_calloc(size_t nmemb, size_t size, void **dst)
{
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, dst, AKERR_NULLPOINTER, "dst=%p", (void *)dst);
*dst = NULL;
FAIL_ZERO_RETURN(e, nmemb, AKERR_VALUE, "zero-size allocation (nmemb=0)");
FAIL_ZERO_RETURN(e, size, AKERR_VALUE, "zero-size allocation (size=0)");
errno = 0;
*dst = calloc(nmemb, size);
FAIL_ZERO_RETURN(e, *dst, AKSL_ERRNO_OR(ENOMEM), "%zu x %zu bytes", nmemb, size);
SUCCEED_RETURN(e);
}
/*
* realloc(3)'s trap, hidden: on failure it returns NULL *and leaves the original
* block valid*, so the near-universal `p = realloc(p, n)` leaks the original
* every time it fails. Here the old pointer goes in and out through the same
* out-param, and is left untouched -- still valid, still the caller's to free --
* whenever an error is raised. TODO.md 3.1.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_realloc(void **ptr, size_t size)
{
void *resized = NULL;
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, ptr, AKERR_NULLPOINTER, "ptr=%p", (void *)ptr);
FAIL_ZERO_RETURN(e, size, AKERR_VALUE,
"zero-size realloc; use aksl_free to release the block");
errno = 0;
resized = realloc(*ptr, size);
FAIL_ZERO_RETURN(e, resized, AKSL_ERRNO_OR(ENOMEM),
"%zu bytes; the original block at %p is still valid", size, *ptr);
*ptr = resized;
SUCCEED_RETURN(e);
}
Version at 0.2.0: complete the wishlist, document it, gate the docs Closes what was left of TODO.md sections 1, 2 and 3, and rewrites that file to hold outstanding items only. The API break gets a minor bump, because pre-1.0 the soname carries MAJOR.MINOR and 0.1 and 0.2 are therefore different ABIs. Five signatures changed and the ato* contract with them; UPGRADING.md is new and lists every one, with the before/after for the cases the compiler cannot warn about. Section 3.1 is finished: reallocarray with the multiplication checked, aligned_alloc and posix_memalign, asprintf/vasprintf, scanf/vscanf. Four functions on that list are deliberately absent rather than missing -- sprintf, strtok, setbuf and perror -- and TODO.md now says which and why, so nobody adds them thinking they were forgotten. Section 1.9, the cross-cutting tests: tests/test_pool.c drives every failure path AKERR_MAX_ARRAY_ERROR + 10 times and checks the pool after each round, because a wrapper that leaks a slot fails a hundred calls later in unrelated code. It also asserts that each error names the function and file it was raised from, which is what catches a FAIL that migrates into a helper during a refactor: status right, message right, origin quietly lying. tests/negative/ two sources that must FAIL to compile, built with -Werror and registered WILL_FAIL. AKERR_NOIGNORE and the format attributes are enforced by the compiler and by nothing else; drop either and every ordinary test still passes. Thread safety is answered rather than tested: the library is not thread-safe and cannot be made so from here, because libakerror's error pool is an unlocked process-global array. README.md says so plainly and TODO.md carries it as the item blocking any future pthread wrappers. Doxygen is configured and gated. All 147 public functions have @brief, a @param each, @throws per status and @return; EXTRACT_ALL is off and WARN_NO_PARAMDOC on, so `cmake --build build --target docs` fails on an undocumented entity. It ran to 0 warnings. The Doxyfile carries no version -- cmake/RunDoxygen.cmake feeds PROJECT_NUMBER in from project(), so that stays the one place a version is written. CI now builds against the submodule it pins instead of also installing libakerror@main and never linking it, adds -Werror, and gains a sanitizer job. The pre-push hook matches, and runs the docs check too. Coverage: 99.5% of lines (1643/1651), 100% of functions (147/147). The eight uncovered lines are each uncovered on purpose and TODO.md says which and why. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 08:00:16 -04:00
/*
* reallocarray(3): a resize whose element count and element size are multiplied
* *with an overflow check*. `realloc(p, n * size)` is a heap overflow waiting for
* an n large enough to wrap, and the multiplication is exactly the place a
* caller does not think to look. Not every platform has it, so the check is done
* here rather than delegated.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_reallocarray(void **ptr, size_t nmemb, size_t size)
{
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, ptr, AKERR_NULLPOINTER, "ptr=%p", (void *)ptr);
FAIL_ZERO_RETURN(e, nmemb, AKERR_VALUE, "zero-size realloc (nmemb=0)");
FAIL_ZERO_RETURN(e, size, AKERR_VALUE, "zero-size realloc (size=0)");
FAIL_NONZERO_RETURN(e, (nmemb > (size_t)-1 / size), AKERR_OUTOFBOUNDS,
"%zu x %zu overflows size_t", nmemb, size);
return aksl_realloc(ptr, nmemb * size);
}
/*
* Aligned allocation. aligned_alloc(3) requires size to be a multiple of the
* alignment and the alignment to be a power of two supported by the
* implementation; violating either is undefined behaviour that usually just
* returns NULL, so both are checked here and reported as what they are.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_aligned_alloc(size_t alignment, size_t size, void **dst)
{
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, dst, AKERR_NULLPOINTER, "dst=%p", (void *)dst);
*dst = NULL;
FAIL_ZERO_RETURN(e, alignment, AKERR_VALUE, "alignment=0");
FAIL_ZERO_RETURN(e, size, AKERR_VALUE, "zero-size allocation");
FAIL_NONZERO_RETURN(e, ((alignment & (alignment - 1)) != 0), AKERR_VALUE,
"alignment %zu is not a power of two", alignment);
FAIL_NONZERO_RETURN(e, (size % alignment != 0), AKERR_VALUE,
"size %zu is not a multiple of alignment %zu", size, alignment);
errno = 0;
*dst = aligned_alloc(alignment, size);
FAIL_ZERO_RETURN(e, *dst, AKSL_ERRNO_OR(ENOMEM),
"%zu bytes aligned to %zu", size, alignment);
SUCCEED_RETURN(e);
}
/*
* posix_memalign(3) is the older interface, and the one that does not require
* size to be a multiple of the alignment. It also breaks the errno convention --
* it *returns* the error number and leaves errno alone -- which is precisely the
* kind of local irregularity a caller mis-handles once and never notices.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_posix_memalign(void **dst, size_t alignment, size_t size)
{
int failed = 0;
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, dst, AKERR_NULLPOINTER, "dst=%p", (void *)dst);
*dst = NULL;
FAIL_ZERO_RETURN(e, size, AKERR_VALUE, "zero-size allocation");
failed = posix_memalign(dst, alignment, size);
if ( failed != 0 ) {
*dst = NULL;
FAIL_RETURN(e, failed, "%zu bytes aligned to %zu", size, alignment);
}
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/*
* free(NULL) is legal C and does nothing. This treats it as an error anyway,
* deliberately and against libc: in a codebase that routes every allocation
* through aksl_malloc, freeing a pointer you believed was live and finding it
* NULL means something upstream did not happen, and silence there is how a
* double-free or a lost allocation gets found three weeks later.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_free(void *ptr)
{
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, ptr, AKERR_NULLPOINTER, "NULL");
free(ptr);
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/*
* Frees and clears in one step, so the pointer cannot be used or freed twice.
* TODO.md 2.2.13 -- aksl_free leaves the caller holding a dangling pointer, and
* "remember to NULL it afterwards" is exactly the discipline this library is
* supposed to make unnecessary rather than merely possible.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_freep(void **ptr)
{
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, ptr, AKERR_NULLPOINTER, "ptr=%p", (void *)ptr);
FAIL_ZERO_RETURN(e, *ptr, AKERR_NULLPOINTER, "*ptr is NULL");
free(*ptr);
*ptr = NULL;
SUCCEED_RETURN(e);
}
/*
* memset(3) and memcpy(3) cannot fail. They return their destination pointer
* unconditionally, so the old FAIL_ZERO_RETURN(e, memset(...), ...) and
* (memcpy(...) == d) checks were dead code that read as though there were a
* failure mode to catch -- TODO.md 2.2.11. What is worth checking is the
* arguments, which is all that is checked now.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_memset(void *s, int c, size_t n)
{
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, s, AKERR_NULLPOINTER, "s=%p", s);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
memset(s, c, n);
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/*
* Overlapping ranges are rejected rather than left as undefined behaviour. This
* is the one place this wrapper is stricter than memcpy(3) instead of merely
* louder: memcpy with overlap is UB that usually appears to work, right up until
* a compiler version or a buffer length changes and it does not. Callers who
* mean to overlap want aksl_memmove, which is what memmove(3) is for.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_memcpy(void *d, const void *s, size_t n)
{
/*
* Through uintptr_t: comparing two pointers that are not into the same
* object is undefined behaviour in C, and "is one inside the other" is
* precisely a comparison of two pointers we do not yet know are related.
*/
uintptr_t dst = (uintptr_t)d;
uintptr_t src = (uintptr_t)s;
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, d, AKERR_NULLPOINTER, "d=%p, s=%p", d, s);
FAIL_ZERO_RETURN(e, s, AKERR_NULLPOINTER, "d=%p, s=%p", d, s);
FAIL_NONZERO_RETURN(e, (n > 0 && dst < src + n && src < dst + n), AKERR_VALUE,
"d=%p and s=%p overlap over %zu bytes; use aksl_memmove",
d, s, n);
memcpy(d, s, n);
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
akerr_ErrorContext AKERR_NOIGNORE *aksl_memmove(void *d, const void *s, size_t n)
{
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, d, AKERR_NULLPOINTER, "d=%p, s=%p", d, s);
FAIL_ZERO_RETURN(e, s, AKERR_NULLPOINTER, "d=%p, s=%p", d, s);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
memmove(d, s, n);
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/*
* memcmp and memchr report their answer through an out-param rather than the
* return value, because the return value is already spoken for by the error
* context. A memchr that finds nothing is not an error -- *dest is NULL and the
* call succeeds -- since "absent" is an ordinary answer to that question.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_memcmp(const void *a, const void *b, size_t n, int *dest)
{
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, a, AKERR_NULLPOINTER, "a=%p, b=%p, dest=%p", a, b, (void *)dest);
FAIL_ZERO_RETURN(e, b, AKERR_NULLPOINTER, "a=%p, b=%p, dest=%p", a, b, (void *)dest);
FAIL_ZERO_RETURN(e, dest, AKERR_NULLPOINTER, "a=%p, b=%p, dest=%p", a, b, (void *)dest);
*dest = memcmp(a, b, n);
SUCCEED_RETURN(e);
}
akerr_ErrorContext AKERR_NOIGNORE *aksl_memchr(const void *s, int c, size_t n, void **dest)
{
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, s, AKERR_NULLPOINTER, "s=%p, dest=%p", s, (void *)dest);
FAIL_ZERO_RETURN(e, dest, AKERR_NULLPOINTER, "s=%p, dest=%p", s, (void *)dest);
*dest = memchr(s, c, n);
SUCCEED_RETURN(e);
}
/*
* pathname and mode are checked. fopen(NULL, ...) is undefined behaviour, and
* this wrapper used to hand both straight through unexamined -- reachable from
* user input in practice, which is why akbasic validates the filename itself
* before calling DLOAD/DSAVE with a comment pointing at TODO.md 2.2.2.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_fopen(
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
const char *pathname,
const char *mode,
FILE **fp)
{
PREPARE_ERROR(e);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
FAIL_ZERO_RETURN(e, fp, AKERR_NULLPOINTER, "pathname=%p, mode=%p, fp=%p",
(void *)pathname, (void *)mode, (void *)fp);
*fp = NULL;
FAIL_ZERO_RETURN(e, pathname, AKERR_NULLPOINTER, "pathname=%p, mode=%p, fp=%p",
(void *)pathname, (void *)mode, (void *)fp);
FAIL_ZERO_RETURN(e, mode, AKERR_NULLPOINTER, "pathname=%p, mode=%p, fp=%p",
(void *)pathname, (void *)mode, (void *)fp);
errno = 0;
*fp = fopen(pathname, mode);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
FAIL_ZERO_RETURN(e, *fp, AKSL_ERRNO_OR(AKERR_IO), "%s", pathname);
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/*
* fread and fwrite both report how much they actually transferred, through a
* required out-param. Three things were wrong with the old pair (TODO.md 2.2.3),
* and the count is the fix for the worst of them: a caller who got AKERR_EOF had
* no way to find out how much data had arrived before the stream ran out, which
* makes the EOF status almost useless for the partial-read case it exists to
* report. *nmemb_out is always written, including on the error paths.
*
* The other two: ptr was never NULL-checked in either function, and a transfer
* that came up short with neither feof nor ferror set fell straight through to
* SUCCEED_RETURN -- a short transfer reported as complete success, which is the
* exact failure mode this library exists to prevent. That case is now AKERR_IO.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_fread(
void *ptr,
size_t size, size_t nmemb,
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
FILE *fp,
size_t *nmemb_out)
{
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
size_t nmemr = 0;
PREPARE_ERROR(e);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
FAIL_ZERO_RETURN(e, nmemb_out, AKERR_NULLPOINTER, "ptr=%p, fp=%p, nmemb_out=%p",
ptr, (void *)fp, (void *)nmemb_out);
*nmemb_out = 0;
FAIL_ZERO_RETURN(e, ptr, AKERR_NULLPOINTER, "ptr=%p, fp=%p, nmemb_out=%p",
ptr, (void *)fp, (void *)nmemb_out);
FAIL_ZERO_RETURN(e, fp, AKERR_NULLPOINTER, "ptr=%p, fp=%p, nmemb_out=%p",
ptr, (void *)fp, (void *)nmemb_out);
errno = 0;
nmemr = fread(ptr, size, nmemb, fp);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
*nmemb_out = nmemr;
if ( nmemr != nmemb ) {
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
FAIL_NONZERO_RETURN(e, ferror(fp), AKSL_ERRNO_OR(AKERR_IO),
"read %zu of %zu members", nmemr, nmemb);
FAIL_NONZERO_RETURN(e, feof(fp), AKERR_EOF,
"EOF after %zu of %zu members", nmemr, nmemb);
FAIL_RETURN(e, AKERR_IO,
"short read: %zu of %zu members, with neither EOF nor a stream error",
nmemr, nmemb);
}
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/*
* ferror before feof, and no feof at all on the write path. The old version
* asked feof() first -- meaningless when writing -- and reported a failed write
* with the message "Error reading file".
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_fwrite(
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
const void *ptr,
size_t size, size_t nmemb,
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
FILE *fp,
size_t *nmemb_out)
{
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
size_t nmemw = 0;
PREPARE_ERROR(e);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
FAIL_ZERO_RETURN(e, nmemb_out, AKERR_NULLPOINTER, "ptr=%p, fp=%p, nmemb_out=%p",
ptr, (void *)fp, (void *)nmemb_out);
*nmemb_out = 0;
FAIL_ZERO_RETURN(e, ptr, AKERR_NULLPOINTER, "ptr=%p, fp=%p, nmemb_out=%p",
ptr, (void *)fp, (void *)nmemb_out);
FAIL_ZERO_RETURN(e, fp, AKERR_NULLPOINTER, "ptr=%p, fp=%p, nmemb_out=%p",
ptr, (void *)fp, (void *)nmemb_out);
errno = 0;
nmemw = fwrite(ptr, size, nmemb, fp);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
*nmemb_out = nmemw;
if ( nmemw != nmemb ) {
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
FAIL_NONZERO_RETURN(e, ferror(fp), AKSL_ERRNO_OR(AKERR_IO),
"wrote %zu of %zu members", nmemw, nmemb);
FAIL_RETURN(e, AKERR_IO,
"short write: %zu of %zu members, with no stream error set",
nmemw, nmemb);
}
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/*
* Closing an already-closed stream is undefined behaviour and there is no way to
* detect it from here -- the FILE * is freed, so even reading it to check is the
* bug. This wrapper cannot help; do not do it.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_fclose(FILE *stream)
{
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, stream, AKERR_NULLPOINTER, "NULL");
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
errno = 0;
FAIL_NONZERO_RETURN(e, fclose(stream), AKSL_ERRNO_OR(AKERR_IO),
"fclose failed, and any buffered data is lost");
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/*
* Formatted output.
*
* The va_list forms below do the work and the variadic forms are thin wrappers
* over them, which is both what §3.1 of TODO.md asked for -- so consumers can
* build their own variadic wrappers -- and what makes the va_end rule below
* checkable in one place instead of three.
*
* Three things every one of these now does that the old versions did not:
*
* - va_end matches every va_start, on the failure path as well as the happy
* one. Omitting it is undefined behaviour per the standard and leaks
* register-save state on some ABIs. akbasic's text sink ran this UB on every
* line of program output without anything visibly misbehaving, which is
* exactly what made it worth fixing before something did. TODO.md 2.1.4.
* - *count is written on every path. It used to be left holding vprintf's -1
* after a failure, so a caller who read the length rather than the status got
* a negative byte count out of a function that had already failed. It is now
* 0 whenever an error is raised.
* - errno is cleared before the call and read back through AKSL_ERRNO_OR, so a
* failure can never be reported with a stale -- or with a zero -- status.
*
* aksl_sprintf is gone. It wrapped vsprintf, which cannot be bounded, and an
* error-handling wrapper around an unbounded write is precisely the sharp edge
* this library exists to remove. aksl_snprintf replaces it, and treats
* truncation as the failure it is rather than as a short success. TODO.md 2.2.4.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_vprintf(int *count, const char *restrict format, va_list args)
{
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
int written = 0;
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, count, AKERR_NULLPOINTER, "count=%p, format=%p", (void *)count, (void *)format);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
*count = 0;
FAIL_ZERO_RETURN(e, format, AKERR_NULLPOINTER, "count=%p, format=%p", (void *)count, (void *)format);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
errno = 0;
written = vprintf(format, args);
FAIL_NONZERO_RETURN(e, (written < 0), AKSL_ERRNO_OR(AKERR_IO), "Short write");
*count = written;
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
akerr_ErrorContext AKERR_NOIGNORE *aksl_printf(int *count, const char *restrict format, ...)
{
va_list args;
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
akerr_ErrorContext *raised = NULL;
va_start(args, format);
raised = aksl_vprintf(count, format, args);
va_end(args);
return raised;
}
akerr_ErrorContext AKERR_NOIGNORE *aksl_vfprintf(int *count, FILE *restrict stream, const char *restrict format, va_list args)
{
int written = 0;
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, count, AKERR_NULLPOINTER, "count=%p, stream=%p, format=%p", (void *)count, (void *)stream, (void *)format);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
*count = 0;
FAIL_ZERO_RETURN(e, stream, AKERR_NULLPOINTER, "count=%p, stream=%p, format=%p", (void *)count, (void *)stream, (void *)format);
FAIL_ZERO_RETURN(e, format, AKERR_NULLPOINTER, "count=%p, stream=%p, format=%p", (void *)count, (void *)stream, (void *)format);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
errno = 0;
written = vfprintf(stream, format, args);
FAIL_NONZERO_RETURN(e, (written < 0), AKSL_ERRNO_OR(AKERR_IO), "Short write");
*count = written;
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
akerr_ErrorContext AKERR_NOIGNORE *aksl_fprintf(int *count, FILE *restrict stream, const char *restrict format, ...)
{
va_list args;
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
akerr_ErrorContext *raised = NULL;
va_start(args, format);
raised = aksl_vfprintf(count, stream, format, args);
va_end(args);
return raised;
}
/*
* Truncation is an error, not a short success. snprintf(3) reports the length it
* *would* have written and silently drops the rest, which is the single most
* common way a bounded write goes wrong unnoticed; a caller who wanted to know
* would have had to compare the return against the buffer size by hand, which is
* the check this library exists to stop people forgetting. *count is the number
* of bytes written excluding the terminating NUL, and is 0 on any failure.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_vsnprintf(int *count, char *restrict str, size_t size, const char *restrict format, va_list args)
{
int needed = 0;
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, count, AKERR_NULLPOINTER, "count=%p, str=%p, format=%p", (void *)count, (void *)str, (void *)format);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
*count = 0;
FAIL_ZERO_RETURN(e, str, AKERR_NULLPOINTER, "count=%p, str=%p, format=%p", (void *)count, (void *)str, (void *)format);
FAIL_ZERO_RETURN(e, format, AKERR_NULLPOINTER, "count=%p, str=%p, format=%p", (void *)count, (void *)str, (void *)format);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
FAIL_ZERO_RETURN(e, size, AKERR_VALUE, "size=0 leaves no room even for the terminating NUL");
errno = 0;
needed = vsnprintf(str, size, format, args);
FAIL_NONZERO_RETURN(e, (needed < 0), AKSL_ERRNO_OR(AKERR_IO), "Output error");
FAIL_NONZERO_RETURN(e, ((size_t)needed >= size), AKERR_OUTOFBOUNDS,
"output truncated: %d bytes needed, %zu available", needed, size);
*count = needed;
SUCCEED_RETURN(e);
}
akerr_ErrorContext AKERR_NOIGNORE *aksl_snprintf(int *count, char *restrict str, size_t size, const char *restrict format, ...)
{
va_list args;
akerr_ErrorContext *raised = NULL;
va_start(args, format);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
raised = aksl_vsnprintf(count, str, size, format, args);
va_end(args);
return raised;
}
Version at 0.2.0: complete the wishlist, document it, gate the docs Closes what was left of TODO.md sections 1, 2 and 3, and rewrites that file to hold outstanding items only. The API break gets a minor bump, because pre-1.0 the soname carries MAJOR.MINOR and 0.1 and 0.2 are therefore different ABIs. Five signatures changed and the ato* contract with them; UPGRADING.md is new and lists every one, with the before/after for the cases the compiler cannot warn about. Section 3.1 is finished: reallocarray with the multiplication checked, aligned_alloc and posix_memalign, asprintf/vasprintf, scanf/vscanf. Four functions on that list are deliberately absent rather than missing -- sprintf, strtok, setbuf and perror -- and TODO.md now says which and why, so nobody adds them thinking they were forgotten. Section 1.9, the cross-cutting tests: tests/test_pool.c drives every failure path AKERR_MAX_ARRAY_ERROR + 10 times and checks the pool after each round, because a wrapper that leaks a slot fails a hundred calls later in unrelated code. It also asserts that each error names the function and file it was raised from, which is what catches a FAIL that migrates into a helper during a refactor: status right, message right, origin quietly lying. tests/negative/ two sources that must FAIL to compile, built with -Werror and registered WILL_FAIL. AKERR_NOIGNORE and the format attributes are enforced by the compiler and by nothing else; drop either and every ordinary test still passes. Thread safety is answered rather than tested: the library is not thread-safe and cannot be made so from here, because libakerror's error pool is an unlocked process-global array. README.md says so plainly and TODO.md carries it as the item blocking any future pthread wrappers. Doxygen is configured and gated. All 147 public functions have @brief, a @param each, @throws per status and @return; EXTRACT_ALL is off and WARN_NO_PARAMDOC on, so `cmake --build build --target docs` fails on an undocumented entity. It ran to 0 warnings. The Doxyfile carries no version -- cmake/RunDoxygen.cmake feeds PROJECT_NUMBER in from project(), so that stays the one place a version is written. CI now builds against the submodule it pins instead of also installing libakerror@main and never linking it, adds -Werror, and gains a sanitizer job. The pre-push hook matches, and runs the docs check too. Coverage: 99.5% of lines (1643/1651), 100% of functions (147/147). The eight uncovered lines are each uncovered on purpose and TODO.md says which and why. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 08:00:16 -04:00
/*
* The allocating form: no buffer, no size, no truncation to worry about. Where
* aksl_snprintf is right for a fixed destination, this is right when the length
* is not knowable in advance and the result is short-lived enough not to want a
* whole aksl_StrBuf.
*
* asprintf(3) is a GNU extension, so it is built here on the two-pass vsnprintf
* measure-then-write rather than called, which keeps this off _GNU_SOURCE. *dest
* is the caller's to release with aksl_free.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_vasprintf(int *count, char **dest, const char *restrict format, va_list args)
{
va_list measure;
int needed = 0;
char *buf = NULL;
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, count, AKERR_NULLPOINTER, "count=%p, dest=%p, format=%p",
(void *)count, (void *)dest, (void *)format);
*count = 0;
FAIL_ZERO_RETURN(e, dest, AKERR_NULLPOINTER, "count=%p, dest=%p, format=%p",
(void *)count, (void *)dest, (void *)format);
*dest = NULL;
FAIL_ZERO_RETURN(e, format, AKERR_NULLPOINTER, "count=%p, dest=%p, format=%p",
(void *)count, (void *)dest, (void *)format);
va_copy(measure, args);
needed = vsnprintf(NULL, 0, format, measure);
va_end(measure);
FAIL_NONZERO_RETURN(e, (needed < 0), AKSL_ERRNO_OR(AKERR_IO), "could not format the arguments");
ATTEMPT {
CATCH(e, aksl_malloc((size_t)needed + 1, (void **)&buf));
/*
* Cannot truncate -- the buffer was sized from this same format and
* arguments -- but the return is checked anyway, because "cannot happen"
* is a claim about the line above rather than about vsnprintf.
*/
FAIL_NONZERO_BREAK(e, (vsnprintf(buf, (size_t)needed + 1, format, args) != needed),
AKERR_IO, "formatted output changed length between passes");
*dest = buf;
*count = needed;
} CLEANUP {
if ( *dest == NULL && buf != NULL ) {
akerr_ErrorContext *released = aksl_free(buf);
if ( released != NULL ) {
released = akerr_release_error(released);
}
}
} PROCESS(e) {
} FINISH(e, true);
SUCCEED_RETURN(e);
}
akerr_ErrorContext AKERR_NOIGNORE *aksl_asprintf(int *count, char **dest, const char *restrict format, ...)
{
va_list args;
akerr_ErrorContext *raised = NULL;
va_start(args, format);
raised = aksl_vasprintf(count, dest, format, args);
va_end(args);
return raised;
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/*
* String to number.
*
* The ato* family used to be the one place in this library where a libc failure
* passed unnoticed: atoi("not a number") handed back success and 0, and
* atoi("99999999999999999999") handed back success and a wrapped value, because
* atoi(3) has no error channel at all. A library whose entire value proposition
* is turning silent libc failures into error contexts cannot ship that.
* TODO.md 2.1.5.
*
* The strto* wrappers below are the real implementation and the ato* wrappers
* are three-line calls into them, which is what TODO.md 3.1 asked for on its own
* account -- akbasic had to hand-write ~60 lines of exactly this (its
* src/convert.c) because the library would not do it.
*
* The contract, for every function here:
*
* no digits consumed -> AKERR_VALUE
* trailing junk, endptr NULL -> AKERR_VALUE
* value outside the type -> ERANGE
* *dest is 0 on every failure
*
* Pass a non-NULL endptr to opt out of the trailing-junk check and parse a
* number out of the front of a longer string; the ato* forms pass NULL, so they
* require the whole string to be a number. Leading whitespace is accepted
* throughout, as strtol(3) accepts it.
*/
/*
* strtoul(3) and strtoull(3) accept a leading '-' and hand back the negation
* wrapped into the unsigned range, so "-1" parses as ULONG_MAX and reports no
* error whatsoever. That is a silent failure of precisely the kind this library
* exists to surface, so the sign is rejected before the call rather than after.
*/
static int has_negative_sign(const char *nptr)
{
while ( isspace((unsigned char)*nptr) ) {
nptr++;
}
return (*nptr == '-') ? 1 : 0;
}
akerr_ErrorContext AKERR_NOIGNORE *aksl_strtol(const char *nptr, char **endptr, int base, long *dest)
{
char *end = NULL;
long value = 0;
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, nptr, AKERR_NULLPOINTER, "nptr=%p, dest=%p", (void *)nptr, (void *)dest);
FAIL_ZERO_RETURN(e, dest, AKERR_NULLPOINTER, "nptr=%p, dest=%p", (void *)nptr, (void *)dest);
*dest = 0;
errno = 0;
value = strtol(nptr, &end, base);
FAIL_NONZERO_RETURN(e, (end == nptr), AKERR_VALUE, "no digits in \"%s\"", nptr);
FAIL_NONZERO_RETURN(e, (errno == ERANGE), ERANGE, "\"%s\" is out of range for long", nptr);
FAIL_NONZERO_RETURN(e, (errno != 0), errno, "\"%s\"", nptr);
if ( endptr != NULL ) {
*endptr = end;
} else {
FAIL_NONZERO_RETURN(e, (*end != '\0'), AKERR_VALUE,
"trailing junk \"%s\" after the number in \"%s\"", end, nptr);
}
*dest = value;
SUCCEED_RETURN(e);
}
2026-05-24 09:52:24 -04:00
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
akerr_ErrorContext AKERR_NOIGNORE *aksl_strtoll(const char *nptr, char **endptr, int base, long long *dest)
2026-05-24 09:52:24 -04:00
{
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
char *end = NULL;
long long value = 0;
2026-05-24 09:52:24 -04:00
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, nptr, AKERR_NULLPOINTER, "nptr=%p, dest=%p", (void *)nptr, (void *)dest);
FAIL_ZERO_RETURN(e, dest, AKERR_NULLPOINTER, "nptr=%p, dest=%p", (void *)nptr, (void *)dest);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
*dest = 0;
errno = 0;
value = strtoll(nptr, &end, base);
FAIL_NONZERO_RETURN(e, (end == nptr), AKERR_VALUE, "no digits in \"%s\"", nptr);
FAIL_NONZERO_RETURN(e, (errno == ERANGE), ERANGE, "\"%s\" is out of range for long long", nptr);
FAIL_NONZERO_RETURN(e, (errno != 0), errno, "\"%s\"", nptr);
if ( endptr != NULL ) {
*endptr = end;
} else {
FAIL_NONZERO_RETURN(e, (*end != '\0'), AKERR_VALUE,
"trailing junk \"%s\" after the number in \"%s\"", end, nptr);
}
*dest = value;
2026-05-24 09:52:24 -04:00
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
akerr_ErrorContext AKERR_NOIGNORE *aksl_strtoul(const char *nptr, char **endptr, int base, unsigned long *dest)
2026-05-24 09:52:24 -04:00
{
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
char *end = NULL;
unsigned long value = 0;
2026-05-24 09:52:24 -04:00
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, nptr, AKERR_NULLPOINTER, "nptr=%p, dest=%p", (void *)nptr, (void *)dest);
FAIL_ZERO_RETURN(e, dest, AKERR_NULLPOINTER, "nptr=%p, dest=%p", (void *)nptr, (void *)dest);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
*dest = 0;
FAIL_NONZERO_RETURN(e, has_negative_sign(nptr), AKERR_VALUE,
"\"%s\" is negative and the destination is unsigned", nptr);
errno = 0;
value = strtoul(nptr, &end, base);
FAIL_NONZERO_RETURN(e, (end == nptr), AKERR_VALUE, "no digits in \"%s\"", nptr);
FAIL_NONZERO_RETURN(e, (errno == ERANGE), ERANGE, "\"%s\" is out of range for unsigned long", nptr);
FAIL_NONZERO_RETURN(e, (errno != 0), errno, "\"%s\"", nptr);
if ( endptr != NULL ) {
*endptr = end;
} else {
FAIL_NONZERO_RETURN(e, (*end != '\0'), AKERR_VALUE,
"trailing junk \"%s\" after the number in \"%s\"", end, nptr);
}
*dest = value;
2026-05-24 09:52:24 -04:00
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
akerr_ErrorContext AKERR_NOIGNORE *aksl_strtoull(const char *nptr, char **endptr, int base, unsigned long long *dest)
2026-05-24 09:52:24 -04:00
{
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
char *end = NULL;
unsigned long long value = 0;
2026-05-24 09:52:24 -04:00
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, nptr, AKERR_NULLPOINTER, "nptr=%p, dest=%p", (void *)nptr, (void *)dest);
FAIL_ZERO_RETURN(e, dest, AKERR_NULLPOINTER, "nptr=%p, dest=%p", (void *)nptr, (void *)dest);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
*dest = 0;
FAIL_NONZERO_RETURN(e, has_negative_sign(nptr), AKERR_VALUE,
"\"%s\" is negative and the destination is unsigned", nptr);
errno = 0;
value = strtoull(nptr, &end, base);
FAIL_NONZERO_RETURN(e, (end == nptr), AKERR_VALUE, "no digits in \"%s\"", nptr);
FAIL_NONZERO_RETURN(e, (errno == ERANGE), ERANGE, "\"%s\" is out of range for unsigned long long", nptr);
FAIL_NONZERO_RETURN(e, (errno != 0), errno, "\"%s\"", nptr);
if ( endptr != NULL ) {
*endptr = end;
} else {
FAIL_NONZERO_RETURN(e, (*end != '\0'), AKERR_VALUE,
"trailing junk \"%s\" after the number in \"%s\"", end, nptr);
}
*dest = value;
2026-05-24 09:52:24 -04:00
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/*
* The floating-point forms report ERANGE for underflow as well as overflow,
* because that is the only signal strtod(3) gives and losing all the significant
* digits of "1e-400" is a conversion failure however it is spelled. "inf",
* "infinity" and "nan" are accepted, in any case, since strtod accepts them and
* they are exact round-trips rather than approximations of something else.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_strtod(const char *nptr, char **endptr, double *dest)
{
char *end = NULL;
double value = 0.0;
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, nptr, AKERR_NULLPOINTER, "nptr=%p, dest=%p", (void *)nptr, (void *)dest);
FAIL_ZERO_RETURN(e, dest, AKERR_NULLPOINTER, "nptr=%p, dest=%p", (void *)nptr, (void *)dest);
*dest = 0.0;
errno = 0;
value = strtod(nptr, &end);
FAIL_NONZERO_RETURN(e, (end == nptr), AKERR_VALUE, "no digits in \"%s\"", nptr);
FAIL_NONZERO_RETURN(e, (errno == ERANGE), ERANGE, "\"%s\" is out of range for double", nptr);
FAIL_NONZERO_RETURN(e, (errno != 0), errno, "\"%s\"", nptr);
if ( endptr != NULL ) {
*endptr = end;
} else {
FAIL_NONZERO_RETURN(e, (*end != '\0'), AKERR_VALUE,
"trailing junk \"%s\" after the number in \"%s\"", end, nptr);
}
*dest = value;
SUCCEED_RETURN(e);
}
akerr_ErrorContext AKERR_NOIGNORE *aksl_strtof(const char *nptr, char **endptr, float *dest)
{
char *end = NULL;
float value = 0.0f;
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, nptr, AKERR_NULLPOINTER, "nptr=%p, dest=%p", (void *)nptr, (void *)dest);
FAIL_ZERO_RETURN(e, dest, AKERR_NULLPOINTER, "nptr=%p, dest=%p", (void *)nptr, (void *)dest);
*dest = 0.0f;
errno = 0;
value = strtof(nptr, &end);
FAIL_NONZERO_RETURN(e, (end == nptr), AKERR_VALUE, "no digits in \"%s\"", nptr);
FAIL_NONZERO_RETURN(e, (errno == ERANGE), ERANGE, "\"%s\" is out of range for float", nptr);
FAIL_NONZERO_RETURN(e, (errno != 0), errno, "\"%s\"", nptr);
if ( endptr != NULL ) {
*endptr = end;
} else {
FAIL_NONZERO_RETURN(e, (*end != '\0'), AKERR_VALUE,
"trailing junk \"%s\" after the number in \"%s\"", end, nptr);
}
*dest = value;
SUCCEED_RETURN(e);
}
akerr_ErrorContext AKERR_NOIGNORE *aksl_strtold(const char *nptr, char **endptr, long double *dest)
2026-05-25 21:29:46 -04:00
{
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
char *end = NULL;
long double value = 0.0L;
2026-05-25 21:29:46 -04:00
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, nptr, AKERR_NULLPOINTER, "nptr=%p, dest=%p", (void *)nptr, (void *)dest);
FAIL_ZERO_RETURN(e, dest, AKERR_NULLPOINTER, "nptr=%p, dest=%p", (void *)nptr, (void *)dest);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
*dest = 0.0L;
errno = 0;
value = strtold(nptr, &end);
FAIL_NONZERO_RETURN(e, (end == nptr), AKERR_VALUE, "no digits in \"%s\"", nptr);
FAIL_NONZERO_RETURN(e, (errno == ERANGE), ERANGE, "\"%s\" is out of range for long double", nptr);
FAIL_NONZERO_RETURN(e, (errno != 0), errno, "\"%s\"", nptr);
if ( endptr != NULL ) {
*endptr = end;
} else {
FAIL_NONZERO_RETURN(e, (*end != '\0'), AKERR_VALUE,
"trailing junk \"%s\" after the number in \"%s\"", end, nptr);
}
*dest = value;
SUCCEED_RETURN(e);
}
/*
* The ato* forms keep their libc-compatible names and signatures but not libc's
* silence: base 10, the whole string must be a number, and a value that does not
* fit the destination type is ERANGE. "0x10" is therefore AKERR_VALUE here --
* base 10 stops at the 'x' and the rest is trailing junk. Reach for
* aksl_strtol(nptr, NULL, 0, &dest) if you want the 0x/0 prefixes honoured.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_atoi(const char *nptr, int *dest)
{
long value = 0;
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, dest, AKERR_NULLPOINTER, "nptr=%p, dest=%p", (void *)nptr, (void *)dest);
*dest = 0;
ATTEMPT {
CATCH(e, aksl_strtol(nptr, NULL, 10, &value));
/*
* strtol already caught anything outside long. This narrows to int, and
* is compiled out where the two are the same width -- leaving it in
* would be a comparison that -Wtype-limits can prove is never true.
*/
#if LONG_MAX > INT_MAX
FAIL_NONZERO_BREAK(e, (value < INT_MIN || value > INT_MAX), ERANGE,
"\"%s\" is out of range for int", nptr);
#endif
*dest = (int)value;
} CLEANUP {
} PROCESS(e) {
} FINISH(e, true);
2026-05-25 21:29:46 -04:00
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
akerr_ErrorContext AKERR_NOIGNORE *aksl_atol(const char *nptr, long *dest)
{
return aksl_strtol(nptr, NULL, 10, dest);
}
akerr_ErrorContext AKERR_NOIGNORE *aksl_atoll(const char *nptr, long long *dest)
{
return aksl_strtoll(nptr, NULL, 10, dest);
}
akerr_ErrorContext AKERR_NOIGNORE *aksl_atof(const char *nptr, double *dest)
{
return aksl_strtod(nptr, NULL, dest);
}
/*
* realpath(3), with the destination buffer made expressible.
*
* The old two-argument form had three separate problems, all of them reachable
* from ordinary use (TODO.md 2.1.6): resolved_path was never NULL-checked, so
* realpath(path, NULL) allocated a buffer the wrapper then discarded and leaked;
* there was no way for a caller to say how big the buffer was, so everyone had
* to know to supply PATH_MAX bytes; and the failure path formatted resolved_path
* with %s even though realpath(3) leaves its contents unspecified on failure --
* so for the normal caller, who passed an uninitialised stack buffer, the error
* path itself read uninitialised memory and could crash.
*
* Taking the length fixes the first two and lets this refuse an undersized
* buffer up front, since realpath(3) offers no way to bound its own write. The
* message names the input path only.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_realpath(const char *restrict path, char *restrict resolved_path, size_t buflen)
2026-05-24 09:52:24 -04:00
{
char *result = NULL;
PREPARE_ERROR(e);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
FAIL_ZERO_RETURN(e, path, AKERR_NULLPOINTER, "path=%p, resolved_path=%p",
(void *)path, (void *)resolved_path);
FAIL_ZERO_RETURN(e, resolved_path, AKERR_NULLPOINTER, "path=%p, resolved_path=%p",
(void *)path, (void *)resolved_path);
FAIL_NONZERO_RETURN(e, (buflen < PATH_MAX), AKERR_OUTOFBOUNDS,
"buflen %zu is below PATH_MAX (%d); realpath(3) cannot be bounded "
"below it, so a shorter buffer cannot be used safely",
buflen, PATH_MAX);
resolved_path[0] = '\0';
errno = 0;
2026-05-24 09:52:24 -04:00
result = realpath(path, resolved_path);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
FAIL_ZERO_RETURN(e, result, AKSL_ERRNO_OR(AKERR_IO), "path=%s", path);
2026-05-24 09:52:24 -04:00
SUCCEED_RETURN(e);
}
2026-06-02 17:11:38 -04:00
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/*
* The allocating variant. realpath(path, NULL) sizes and allocates the buffer
* itself, which is the right answer when the caller has no PATH_MAX-sized buffer
* to hand; *dest is the caller's to release with aksl_free.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_realpath_alloc(const char *restrict path, char **dest)
2026-06-02 17:11:38 -04:00
{
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
char *result = NULL;
2026-06-02 17:11:38 -04:00
PREPARE_ERROR(e);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
FAIL_ZERO_RETURN(e, path, AKERR_NULLPOINTER, "path=%p, dest=%p", (void *)path, (void *)dest);
FAIL_ZERO_RETURN(e, dest, AKERR_NULLPOINTER, "path=%p, dest=%p", (void *)path, (void *)dest);
*dest = NULL;
errno = 0;
result = realpath(path, NULL);
FAIL_ZERO_RETURN(e, result, AKSL_ERRNO_OR(AKERR_IO), "path=%s", path);
*dest = result;
SUCCEED_RETURN(e);
}
/*
* djb2 over an explicit length, so embedded NUL bytes hash like any other byte.
*
* The cursor is an unsigned char *, not a char *. On x86 and ARM Linux plain
* char is signed, so every byte >= 0x80 used to contribute a sign-extended
* negative value: the hash differed from canonical djb2, and -- worse -- differed
* between platforms depending on the signedness of char. Benign for the 7-bit
* ASCII identifiers akbasic hashes, and quietly wrong for the first caller to
* key a table on a filename or a UTF-8 string literal. TODO.md 2.2.6.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_strhash_djb2(const char *str, size_t len, uint32_t *hashval)
{
const unsigned char *cursor = NULL;
2026-06-02 17:11:38 -04:00
uint32_t h = 5381;
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, str, AKERR_NULLPOINTER, "str=%p, hashval=%p", (void *)str, (void *)hashval);
FAIL_ZERO_RETURN(e, hashval, AKERR_NULLPOINTER, "str=%p, hashval=%p", (void *)str, (void *)hashval);
cursor = (const unsigned char *)str;
2026-06-02 17:11:38 -04:00
while (len--) {
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
h = ((h << 5) + h) + *cursor++;
2026-06-02 17:11:38 -04:00
}
*hashval = h;
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/* The NUL-terminated convenience form TODO.md 3.6 asked for. */
akerr_ErrorContext AKERR_NOIGNORE *aksl_strhash_djb2_str(const char *str, uint32_t *hashval)
{
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, str, AKERR_NULLPOINTER, "str=%p, hashval=%p", (void *)str, (void *)hashval);
return aksl_strhash_djb2(str, strlen(str), hashval);
}
akerr_ErrorContext AKERR_NOIGNORE *aksl_list_append(aksl_ListNode *list, aksl_ListNode *obj)
{
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, list, AKERR_NULLPOINTER, "list");
FAIL_ZERO_RETURN(e, obj, AKERR_NULLPOINTER, "obj");
aksl_ListNode *slow = list;
aksl_ListNode *fast = list;
aksl_ListNode *tail = list;
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/*
* Two separate walks, deliberately. The Floyd pass answers "is this list
* finite"; it says nothing about where the tail is, because `slow` stops at
* the midpoint. Conflating the two is what made this function truncate every
* list of two or more nodes -- see TODO.md 2.1.1.
*/
while ( fast != NULL && fast->next != NULL ) {
slow = slow->next;
fast = fast->next->next;
if ( fast == slow) {
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
FAIL_RETURN(e, AKERR_CIRCULAR_REFERENCE, "%p", (void *)list);
}
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/*
* Only now is the list known to terminate, so this walk is guaranteed to.
* It doubles as the aliasing check: appending a node that is already in the
* list would relink it behind itself and orphan everything between its old
* position and the tail, so it is refused. The walk was happening anyway,
* which is what makes the check free.
*/
FAIL_NONZERO_RETURN(e, (tail == obj), AKERR_VALUE,
"obj %p is already the head of this list", (void *)obj);
while ( tail->next != NULL ) {
tail = tail->next;
FAIL_NONZERO_RETURN(e, (tail == obj), AKERR_VALUE,
"obj %p is already in this list", (void *)obj);
}
tail->next = obj;
obj->next = NULL;
obj->prev = tail;
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/*
* `head` is required, not optional. Popping the head node used to leave the
* caller's own head pointer aimed at a node that is no longer in the list, and
* there was no way for the caller to learn the new one -- TODO.md 2.2.12. Taking
* the head by reference makes the correct call the only call that compiles.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_list_pop(aksl_ListNode **head, aksl_ListNode *node)
{
PREPARE_ERROR(e);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
FAIL_ZERO_RETURN(e, head, AKERR_NULLPOINTER, "head=%p, node=%p", (void *)head, (void *)node);
FAIL_ZERO_RETURN(e, node, AKERR_NULLPOINTER, "head=%p, node=%p", (void *)head, (void *)node);
if ( node->prev != NULL ) {
node->prev->next = node->next;
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
if ( node->next != NULL ) {
node->next->prev = node->prev;
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
if ( *head == node ) {
*head = node->next;
}
node->next = NULL;
node->prev = NULL;
SUCCEED_RETURN(e);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/*
* Zeroing initialisers. Every caller previously had to remember to memset a node
* before its first use, because append and iterate both read next/prev -- and a
* stack-allocated node that skipped it walked into garbage. TODO.md 2.2.14.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_list_node_init(aksl_ListNode *node, void *data)
{
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, node, AKERR_NULLPOINTER, "node");
node->data = data;
node->next = NULL;
node->prev = NULL;
SUCCEED_RETURN(e);
}
akerr_ErrorContext AKERR_NOIGNORE *aksl_tree_node_init(aksl_TreeNode *node, void *leaf)
{
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, node, AKERR_NULLPOINTER, "node");
node->parent = NULL;
node->left = NULL;
node->right = NULL;
node->leaf = leaf;
SUCCEED_RETURN(e);
}
Version at 0.2.0: complete the wishlist, document it, gate the docs Closes what was left of TODO.md sections 1, 2 and 3, and rewrites that file to hold outstanding items only. The API break gets a minor bump, because pre-1.0 the soname carries MAJOR.MINOR and 0.1 and 0.2 are therefore different ABIs. Five signatures changed and the ato* contract with them; UPGRADING.md is new and lists every one, with the before/after for the cases the compiler cannot warn about. Section 3.1 is finished: reallocarray with the multiplication checked, aligned_alloc and posix_memalign, asprintf/vasprintf, scanf/vscanf. Four functions on that list are deliberately absent rather than missing -- sprintf, strtok, setbuf and perror -- and TODO.md now says which and why, so nobody adds them thinking they were forgotten. Section 1.9, the cross-cutting tests: tests/test_pool.c drives every failure path AKERR_MAX_ARRAY_ERROR + 10 times and checks the pool after each round, because a wrapper that leaks a slot fails a hundred calls later in unrelated code. It also asserts that each error names the function and file it was raised from, which is what catches a FAIL that migrates into a helper during a refactor: status right, message right, origin quietly lying. tests/negative/ two sources that must FAIL to compile, built with -Werror and registered WILL_FAIL. AKERR_NOIGNORE and the format attributes are enforced by the compiler and by nothing else; drop either and every ordinary test still passes. Thread safety is answered rather than tested: the library is not thread-safe and cannot be made so from here, because libakerror's error pool is an unlocked process-global array. README.md says so plainly and TODO.md carries it as the item blocking any future pthread wrappers. Doxygen is configured and gated. All 147 public functions have @brief, a @param each, @throws per status and @return; EXTRACT_ALL is off and WARN_NO_PARAMDOC on, so `cmake --build build --target docs` fails on an undocumented entity. It ran to 0 warnings. The Doxyfile carries no version -- cmake/RunDoxygen.cmake feeds PROJECT_NUMBER in from project(), so that stays the one place a version is written. CI now builds against the submodule it pins instead of also installing libakerror@main and never linking it, adds -Werror, and gains a sanitizer job. The pre-push hook matches, and runs the docs check too. Coverage: 99.5% of lines (1643/1651), 100% of functions (147/147). The eight uncovered lines are each uncovered on purpose and TODO.md says which and why. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 08:00:16 -04:00
/** @cond INTERNAL */
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/*
* One entry in the breadth-first traversal queue. Internal, and deliberately not
* aksl_ListNode: the walk needs to carry each node's depth alongside it, which is
* how a cyclic tree is stopped without a visited-set. The caller's lalloc/lfree
* allocate and release these -- which is what they were always documented to do
* and never actually did (TODO.md 2.2.8).
*/
typedef struct TreeQueueEntry {
struct TreeQueueEntry *next;
aksl_TreeNode *node;
int depth;
} TreeQueueEntry;
Version at 0.2.0: complete the wishlist, document it, gate the docs Closes what was left of TODO.md sections 1, 2 and 3, and rewrites that file to hold outstanding items only. The API break gets a minor bump, because pre-1.0 the soname carries MAJOR.MINOR and 0.1 and 0.2 are therefore different ABIs. Five signatures changed and the ato* contract with them; UPGRADING.md is new and lists every one, with the before/after for the cases the compiler cannot warn about. Section 3.1 is finished: reallocarray with the multiplication checked, aligned_alloc and posix_memalign, asprintf/vasprintf, scanf/vscanf. Four functions on that list are deliberately absent rather than missing -- sprintf, strtok, setbuf and perror -- and TODO.md now says which and why, so nobody adds them thinking they were forgotten. Section 1.9, the cross-cutting tests: tests/test_pool.c drives every failure path AKERR_MAX_ARRAY_ERROR + 10 times and checks the pool after each round, because a wrapper that leaks a slot fails a hundred calls later in unrelated code. It also asserts that each error names the function and file it was raised from, which is what catches a FAIL that migrates into a helper during a refactor: status right, message right, origin quietly lying. tests/negative/ two sources that must FAIL to compile, built with -Werror and registered WILL_FAIL. AKERR_NOIGNORE and the format attributes are enforced by the compiler and by nothing else; drop either and every ordinary test still passes. Thread safety is answered rather than tested: the library is not thread-safe and cannot be made so from here, because libakerror's error pool is an unlocked process-global array. README.md says so plainly and TODO.md carries it as the item blocking any future pthread wrappers. Doxygen is configured and gated. All 147 public functions have @brief, a @param each, @throws per status and @return; EXTRACT_ALL is off and WARN_NO_PARAMDOC on, so `cmake --build build --target docs` fails on an undocumented entity. It ran to 0 warnings. The Doxyfile carries no version -- cmake/RunDoxygen.cmake feeds PROJECT_NUMBER in from project(), so that stays the one place a version is written. CI now builds against the submodule it pins instead of also installing libakerror@main and never linking it, adds -Werror, and gains a sanitizer job. The pre-push hook matches, and runs the docs check too. Coverage: 99.5% of lines (1643/1651), 100% of functions (147/147). The eight uncovered lines are each uncovered on purpose and TODO.md says which and why. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 08:00:16 -04:00
/** @endcond */
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/* Append one tree node to the tail of the traversal queue. */
static akerr_ErrorContext AKERR_NOIGNORE *tree_queue_push(
TreeQueueEntry **head,
TreeQueueEntry **tail,
aksl_AllocFunc lalloc,
aksl_TreeNode *node,
int depth)
{
TreeQueueEntry *entry = NULL;
PREPARE_ERROR(e);
ATTEMPT {
CATCH(e, lalloc(sizeof(TreeQueueEntry), (void **)&entry));
FAIL_ZERO_BREAK(e, entry, AKERR_NULLPOINTER,
"allocator returned success but no memory");
entry->next = NULL;
entry->node = node;
entry->depth = depth;
if ( *tail == NULL ) {
*head = entry;
} else {
(*tail)->next = entry;
}
*tail = entry;
} CLEANUP {
} PROCESS(e) {
} FINISH(e, true);
SUCCEED_RETURN(e);
}
/*
* Release every entry still queued. Called from the CLEANUP block of the walk,
* so it runs on the error and break paths as well as the successful one. An
* lfree that itself fails cannot be allowed to abandon the rest of the queue, so
* the first such error is remembered and returned only once the queue is empty.
*/
static akerr_ErrorContext AKERR_NOIGNORE *tree_queue_drain(
TreeQueueEntry **head,
TreeQueueEntry **tail,
aksl_FreeFunc lfree)
{
TreeQueueEntry *entry = NULL;
akerr_ErrorContext *first = NULL;
akerr_ErrorContext *this = NULL;
PREPARE_ERROR(e);
while ( *head != NULL ) {
entry = *head;
*head = entry->next;
this = lfree(entry);
if ( this != NULL && first == NULL ) {
first = this;
} else if ( this != NULL ) {
this = akerr_release_error(this);
}
}
*tail = NULL;
if ( first != NULL ) {
return first;
}
SUCCEED_RETURN(e);
}
/*
* Breadth-first walk. Consumes the queue it is given; on any error it returns
* with the remainder still queued, and the caller's CLEANUP drains it.
*
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
* AKSL_TREE_SEARCH_BFS enqueues left before right, AKSL_TREE_SEARCH_BFS_RIGHT
* the other way round. A tree whose child points back at an ancestor grows the
* recorded depth without bound rather than looping forever, so it is caught by
* the depth cap -- as AKERR_OUTOFBOUNDS, not AKERR_CIRCULAR_REFERENCE, because
* unlike the depth-first walk this one has no ancestor chain to compare against.
*/
static akerr_ErrorContext AKERR_NOIGNORE *tree_bfs_walk(
TreeQueueEntry **head,
TreeQueueEntry **tail,
aksl_TreeNodeIterator iter,
aksl_AllocFunc lalloc,
aksl_FreeFunc lfree,
uint8_t searchmode,
void *data)
{
TreeQueueEntry *entry = NULL;
aksl_TreeNode *node = NULL;
aksl_TreeNode *first = NULL;
aksl_TreeNode *second = NULL;
int depth = 0;
PREPARE_ERROR(e);
while ( *head != NULL ) {
entry = *head;
*head = entry->next;
if ( *head == NULL ) {
*tail = NULL;
}
node = entry->node;
depth = entry->depth;
/*
* Release the entry before doing anything that can fail, so the queue
* the caller's CLEANUP drains never contains an entry we already
* detached from it.
*/
PASS(e, lfree(entry));
FAIL_NONZERO_RETURN(e, (depth >= AKSL_TREE_MAX_DEPTH), AKERR_OUTOFBOUNDS,
"tree deeper than AKSL_TREE_MAX_DEPTH (%d) at node %p",
AKSL_TREE_MAX_DEPTH, (void *)node);
PASS(e, iter(node, data));
if ( searchmode == AKSL_TREE_SEARCH_BFS_RIGHT ) {
first = node->right;
second = node->left;
} else {
first = node->left;
second = node->right;
}
if ( first != NULL ) {
PASS(e, tree_queue_push(head, tail, lalloc, first, depth + 1));
}
if ( second != NULL ) {
PASS(e, tree_queue_push(head, tail, lalloc, second, depth + 1));
}
}
SUCCEED_RETURN(e);
}
/*
* The depth-first recursion.
*
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
* Split out from aksl_tree_iterate so that AKERR_ITERATOR_BREAK *propagates*.
* This frame has no HANDLE block for it, so a break raised by the callback
* unwinds every frame up to the public entry point, which swallows it exactly
* once. In the old single-function form the frame that raised the break handled
* it and returned success, the parent's PASS therefore saw nothing wrong, and
* the walk carried on into the sibling subtree -- TODO.md 2.1.3.
*
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
* `path` is the chain of ancestors of `root` and `depth` its length. Checking
* each node against its own ancestry turns a tree that loops back on itself from
* infinite recursion into AKERR_CIRCULAR_REFERENCE, and the depth cap catches the
* degenerate chain that is merely too deep to recurse over (TODO.md 2.2.7).
*/
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
static akerr_ErrorContext AKERR_NOIGNORE *tree_dfs_walk(
aksl_TreeNode *root,
aksl_TreeNodeIterator iter,
uint8_t searchmode,
void *data,
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
aksl_TreeNode **path,
int depth)
{
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
int i = 0;
PREPARE_ERROR(e);
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
FAIL_NONZERO_RETURN(e, (depth >= AKSL_TREE_MAX_DEPTH), AKERR_OUTOFBOUNDS,
"tree deeper than AKSL_TREE_MAX_DEPTH (%d) at node %p",
AKSL_TREE_MAX_DEPTH, (void *)root);
for ( i = 0; i < depth; i++ ) {
FAIL_NONZERO_RETURN(e, (path[i] == root), AKERR_CIRCULAR_REFERENCE,
"node %p is its own ancestor at depth %d",
(void *)root, depth);
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
path[depth] = root;
2026-06-27 13:13:17 -04:00
ATTEMPT {
switch ( searchmode ) {
case AKSL_TREE_SEARCH_DFS_PREORDER:
CATCH(e, iter(root, data));
if ( root->left != NULL ) {
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
PASS(e, tree_dfs_walk(root->left, iter, searchmode, data, path, depth + 1));
2026-06-27 13:13:17 -04:00
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
if ( root->right != NULL ) {
PASS(e, tree_dfs_walk(root->right, iter, searchmode, data, path, depth + 1));
2026-06-27 13:13:17 -04:00
}
break;
case AKSL_TREE_SEARCH_DFS_POSTORDER:
if ( root->left != NULL ) {
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
PASS(e, tree_dfs_walk(root->left, iter, searchmode, data, path, depth + 1));
2026-06-27 13:13:17 -04:00
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
if ( root->right != NULL ) {
PASS(e, tree_dfs_walk(root->right, iter, searchmode, data, path, depth + 1));
2026-06-27 13:13:17 -04:00
}
CATCH(e, iter(root, data));
break;
case AKSL_TREE_SEARCH_DFS_INORDER:
if ( root->left != NULL ) {
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
PASS(e, tree_dfs_walk(root->left, iter, searchmode, data, path, depth + 1));
2026-06-27 13:13:17 -04:00
}
CATCH(e, iter(root, data));
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
if ( root->right != NULL ) {
PASS(e, tree_dfs_walk(root->right, iter, searchmode, data, path, depth + 1));
2026-06-27 13:13:17 -04:00
}
break;
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
case AKSL_TREE_SEARCH_VISIT:
/* Visit this node and stop; the children are deliberately not walked. */
CATCH(e, iter(root, data));
2026-06-27 13:13:17 -04:00
break;
}
2026-06-27 13:13:17 -04:00
} CLEANUP {
} PROCESS(e) {
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
} FINISH(e, true);
SUCCEED_RETURN(e);
}
Version at 0.2.0: complete the wishlist, document it, gate the docs Closes what was left of TODO.md sections 1, 2 and 3, and rewrites that file to hold outstanding items only. The API break gets a minor bump, because pre-1.0 the soname carries MAJOR.MINOR and 0.1 and 0.2 are therefore different ABIs. Five signatures changed and the ato* contract with them; UPGRADING.md is new and lists every one, with the before/after for the cases the compiler cannot warn about. Section 3.1 is finished: reallocarray with the multiplication checked, aligned_alloc and posix_memalign, asprintf/vasprintf, scanf/vscanf. Four functions on that list are deliberately absent rather than missing -- sprintf, strtok, setbuf and perror -- and TODO.md now says which and why, so nobody adds them thinking they were forgotten. Section 1.9, the cross-cutting tests: tests/test_pool.c drives every failure path AKERR_MAX_ARRAY_ERROR + 10 times and checks the pool after each round, because a wrapper that leaks a slot fails a hundred calls later in unrelated code. It also asserts that each error names the function and file it was raised from, which is what catches a FAIL that migrates into a helper during a refactor: status right, message right, origin quietly lying. tests/negative/ two sources that must FAIL to compile, built with -Werror and registered WILL_FAIL. AKERR_NOIGNORE and the format attributes are enforced by the compiler and by nothing else; drop either and every ordinary test still passes. Thread safety is answered rather than tested: the library is not thread-safe and cannot be made so from here, because libakerror's error pool is an unlocked process-global array. README.md says so plainly and TODO.md carries it as the item blocking any future pthread wrappers. Doxygen is configured and gated. All 147 public functions have @brief, a @param each, @throws per status and @return; EXTRACT_ALL is off and WARN_NO_PARAMDOC on, so `cmake --build build --target docs` fails on an undocumented entity. It ran to 0 warnings. The Doxyfile carries no version -- cmake/RunDoxygen.cmake feeds PROJECT_NUMBER in from project(), so that stays the one place a version is written. CI now builds against the submodule it pins instead of also installing libakerror@main and never linking it, adds -Werror, and gains a sanitizer job. The pre-push hook matches, and runs the docs check too. Coverage: 99.5% of lines (1643/1651), 100% of functions (147/147). The eight uncovered lines are each uncovered on purpose and TODO.md says which and why. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 08:00:16 -04:00
/*
* Public entry point for every tree walk. The contract -- every status, every
* argument -- is documented on the declaration in include/akstdlib.h; what is
* here is why the code looks the way it does.
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_tree_iterate(
aksl_TreeNode *root,
aksl_TreeNodeIterator iter,
aksl_AllocFunc lalloc,
aksl_FreeFunc lfree,
uint8_t searchmode,
void *data)
{
TreeQueueEntry *head = NULL;
TreeQueueEntry *tail = NULL;
/*
* The ancestor chain for the depth-first walk, held here rather than spread
* across the recursion so that a walk costs one bounded allocation-free
* buffer no matter how the tree is shaped.
*/
aksl_TreeNode *path[AKSL_TREE_MAX_DEPTH];
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, root, AKERR_NULLPOINTER, "root=%p, iter=%p", (void *)root, (void *)iter);
FAIL_ZERO_RETURN(e, iter, AKERR_NULLPOINTER, "root=%p, iter=%p", (void *)root, (void *)iter);
if ( lalloc == NULL ) {
lalloc = &aksl_malloc;
}
if ( lfree == NULL ) {
lfree = &aksl_free;
}
/*
* Validate the mode here, outside the ATTEMPT block, so that the walkers can
* assume it. The switch used to have no default at all, so searchmode 99 --
* and AKSL_TREE_SEARCH_VISIT, which the header documented but nothing
* implemented -- fell straight through to success having visited nothing
* (TODO.md 2.2.9).
*/
switch ( searchmode ) {
case AKSL_TREE_SEARCH_DFS_PREORDER:
case AKSL_TREE_SEARCH_DFS_INORDER:
case AKSL_TREE_SEARCH_DFS_POSTORDER:
case AKSL_TREE_SEARCH_VISIT:
case AKSL_TREE_SEARCH_BFS:
case AKSL_TREE_SEARCH_BFS_RIGHT:
break;
default:
FAIL_RETURN(e, AKERR_VALUE, "unknown searchmode %d", searchmode);
}
ATTEMPT {
if ( searchmode == AKSL_TREE_SEARCH_BFS || searchmode == AKSL_TREE_SEARCH_BFS_RIGHT ) {
CATCH(e, tree_queue_push(&head, &tail, lalloc, root, 0));
CATCH(e, tree_bfs_walk(&head, &tail, iter, lalloc, lfree, searchmode, data));
} else {
CATCH(e, tree_dfs_walk(root, iter, searchmode, data, path, 0));
}
} CLEANUP {
/*
* Drains on every path, which is the whole reason the walk above leaves
* the queue alone when it fails. A failure to release the queue must not
* mask the error that got us here, so it is logged and dropped -- but
* dropped by hand rather than with IGNORE(), which logs the context and
* then never releases it back to the pool.
*/
akerr_ErrorContext *drained = tree_queue_drain(&head, &tail, lfree);
if ( drained != NULL ) {
LOG_ERROR_WITH_MESSAGE(drained, "** traversal queue drain failed **");
drained = akerr_release_error(drained);
}
} PROCESS(e) {
2026-06-27 13:13:17 -04:00
} HANDLE(e, AKERR_ITERATOR_BREAK) {
// This is not an error condition, it's just telling us to stop early
SUCCEED_RETURN(e);
} FINISH(e, true);
SUCCEED_RETURN(e);
}
Version at 0.2.0: complete the wishlist, document it, gate the docs Closes what was left of TODO.md sections 1, 2 and 3, and rewrites that file to hold outstanding items only. The API break gets a minor bump, because pre-1.0 the soname carries MAJOR.MINOR and 0.1 and 0.2 are therefore different ABIs. Five signatures changed and the ato* contract with them; UPGRADING.md is new and lists every one, with the before/after for the cases the compiler cannot warn about. Section 3.1 is finished: reallocarray with the multiplication checked, aligned_alloc and posix_memalign, asprintf/vasprintf, scanf/vscanf. Four functions on that list are deliberately absent rather than missing -- sprintf, strtok, setbuf and perror -- and TODO.md now says which and why, so nobody adds them thinking they were forgotten. Section 1.9, the cross-cutting tests: tests/test_pool.c drives every failure path AKERR_MAX_ARRAY_ERROR + 10 times and checks the pool after each round, because a wrapper that leaks a slot fails a hundred calls later in unrelated code. It also asserts that each error names the function and file it was raised from, which is what catches a FAIL that migrates into a helper during a refactor: status right, message right, origin quietly lying. tests/negative/ two sources that must FAIL to compile, built with -Werror and registered WILL_FAIL. AKERR_NOIGNORE and the format attributes are enforced by the compiler and by nothing else; drop either and every ordinary test still passes. Thread safety is answered rather than tested: the library is not thread-safe and cannot be made so from here, because libakerror's error pool is an unlocked process-global array. README.md says so plainly and TODO.md carries it as the item blocking any future pthread wrappers. Doxygen is configured and gated. All 147 public functions have @brief, a @param each, @throws per status and @return; EXTRACT_ALL is off and WARN_NO_PARAMDOC on, so `cmake --build build --target docs` fails on an undocumented entity. It ran to 0 warnings. The Doxyfile carries no version -- cmake/RunDoxygen.cmake feeds PROJECT_NUMBER in from project(), so that stays the one place a version is written. CI now builds against the submodule it pins instead of also installing libakerror@main and never linking it, adds -Werror, and gains a sanitizer job. The pre-push hook matches, and runs the docs check too. Coverage: 99.5% of lines (1643/1651), 100% of functions (147/147). The eight uncovered lines are each uncovered on purpose and TODO.md says which and why. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 08:00:16 -04:00
/*
* Contract documented on the declaration in include/akstdlib.h.
*/
akerr_ErrorContext AKERR_NOIGNORE *aksl_list_iterate(aksl_ListNode *list, aksl_ListNodeIterator iter, void *data)
{
PREPARE_ERROR(e);
FAIL_ZERO_RETURN(e, list, AKERR_NULLPOINTER, "list");
FAIL_ZERO_RETURN(e, iter, AKERR_NULLPOINTER, "iter");
aksl_ListNode *slow = list;
aksl_ListNode *fast = list;
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
aksl_ListNode *node = list;
while ( fast != NULL && fast->next != NULL ) {
slow = slow->next;
fast = fast->next->next;
if ( fast == slow) {
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
FAIL_RETURN(e, AKERR_CIRCULAR_REFERENCE, "%p", (void *)list);
}
}
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
/*
* From `list`, not from `slow`. The Floyd pass above leaves `slow` at the
* midpoint, and starting the visit there skipped the whole first half of the
* list including the head -- see TODO.md 2.1.2.
*/
while ( node != NULL ) {
ATTEMPT {
Fix the six confirmed defects and close the API contract gaps TODO.md section 2.1 recorded six defects reproduced against the built library, and section 2.2 seventeen contract gaps. Both are closed. The four tests registered in AKSL_KNOWN_FAILING_TESTS are folded back into the tests for the things they test, and that list is now empty. The defects: 2.1.1 aksl_list_append conflated Floyd cycle detection with finding the tail, so `tail` tracked the node behind the midpoint. Any append to a list of 2+ nodes silently dropped everything after it. Two separate walks now: Floyd to prove the list is finite, then a plain walk to the end. 2.1.2 aksl_list_iterate started visiting from Floyd's `slow` cursor, so the whole first half of the list -- head included -- was never passed to the callback. It starts at the head. 2.1.3 AKERR_ITERATOR_BREAK did not stop a tree traversal: the frame that raised it handled it and returned success, so the parent carried on into the sibling subtree. The recursion is split out and propagates the break; only the public entry swallows it. 2.1.4 va_end now matches every va_start on every path. 2.1.5 The ato* family had no error channel at all. Reimplemented over a new strto* family with errno cleared, an endptr check and a range check: AKERR_VALUE for junk, ERANGE for overflow. 2.1.6 aksl_realpath never checked resolved_path, could not be told the buffer size, and formatted an unspecified buffer with %s on its own error path. It takes a length; aksl_realpath_alloc is the allocating form. The contract gaps, in brief: errno is cleared before every wrapped call and read back through a fallback so no error can carry status 0; fopen validates pathname and mode; fread/fwrite report the transferred count through a required out-param and no longer call a short transfer a success; aksl_sprintf is gone in favour of aksl_snprintf, which treats truncation as an error; the variadic wrappers carry format attributes; djb2 reads bytes as unsigned; tree traversal is depth- and cycle-bounded and implements BFS, so lalloc/lfree are used rather than merely stored; an unknown searchmode is AKERR_VALUE rather than silent success; list_pop takes the head by reference; aksl_freep, the node initialisers and extern "C" are new. Build: -pg is out of the default build (it never reached the C compiler anyway, and it is what produced the stray gmon.out), -Wall -Wextra are in, and there is a .gitignore. Tests: 11 binaries, all green under the normal and sanitizer builds. Visit-order assertions replace the step counts that could not tell the three depth-first orders apart. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-31 07:14:11 -04:00
CATCH(e, iter(node, data));
node = node->next;
} CLEANUP {
} PROCESS(e) {
} HANDLE(e, AKERR_ITERATOR_BREAK) {
// This is not an error condition, it's just telling us to stop early
SUCCEED_RETURN(e);
} FINISH(e, true);
}
SUCCEED_RETURN(e);
}