Files
libakerror/src/error.c

513 lines
19 KiB
C
Raw Normal View History

#include "akerror.h"
#if defined(AKERR_USE_STDLIB) && AKERR_USE_STDLIB == 1
#include <stdlib.h>
#include <stdarg.h>
#include <stdio.h>
#endif // AKERR_USE_STDLIB
2025-07-20 21:40:14 -04:00
akerr_ErrorContext __akerr_last_ditch;
akerr_ErrorContext *__akerr_last_ignored;
akerr_ErrorUnhandledErrorHandler akerr_handler_unhandled_error;
akerr_ErrorLogFunction akerr_log_method = NULL;
2025-07-20 21:40:14 -04:00
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
/*
* Status-name registry.
*
* Storage is an open-addressed hash table keyed by status value. Both sizes are
* private to this translation unit -- they are deliberately NOT in the public
* header, because a consumer-visible table bound is exactly the ABI hazard this
* registry replaced. Overriding them changes only this file, so a library and
* its consumers can never disagree about the layout.
*
* The table is never resized or rehashed, so a pointer handed out by
* akerr_name_for_status() stays valid for the life of the process. Entries are
* never removed, so probing needs no tombstones.
*/
#ifndef AKERR_STATUS_NAME_SLOTS
#define AKERR_STATUS_NAME_SLOTS 4096
#endif
#ifndef AKERR_MAX_RESERVED_STATUS_RANGES
#define AKERR_MAX_RESERVED_STATUS_RANGES 64
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
#endif
/* Probing masks with SLOTS-1, so the slot count must be a power of two. This is
* the C99-portable spelling of a static assertion (a negative array bound). */
typedef char akerr_assert_name_slots_pow2[
(AKERR_STATUS_NAME_SLOTS > 0 &&
(AKERR_STATUS_NAME_SLOTS & (AKERR_STATUS_NAME_SLOTS - 1)) == 0) ? 1 : -1];
/* Cap occupancy at 75% so linear probing always meets an empty slot. */
#define AKERR_MAX_REGISTERED_STATUS_NAMES \
(AKERR_STATUS_NAME_SLOTS - (AKERR_STATUS_NAME_SLOTS / 4))
#define AKERR_MAX_STATUS_RANGE_OWNER_LENGTH 64
typedef struct
{
int status;
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
int used;
char name[AKERR_MAX_ERROR_NAME_LENGTH];
} akerr_StatusName;
typedef struct
{
int first;
int last;
char owner[AKERR_MAX_STATUS_RANGE_OWNER_LENGTH];
} akerr_StatusRange;
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
static akerr_StatusName akerr_status_names[AKERR_STATUS_NAME_SLOTS];
static int akerr_status_name_count;
static akerr_StatusRange akerr_status_ranges[AKERR_MAX_RESERVED_STATUS_RANGES];
static int akerr_status_range_count;
2025-07-20 21:40:14 -04:00
akerr_ErrorContext AKERR_ARRAY_ERROR[AKERR_MAX_ARRAY_ERROR];
2025-07-20 21:40:14 -04:00
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
/*
* Bounded copy into a fixed buffer. Every argument is checked: this writes
* through a caller-supplied pointer for a caller-supplied length, so a NULL or
* a non-positive capacity here is a memory error waiting to happen, not
* something to absorb and return from quietly.
*
* Exported under the internal __akerr_ prefix rather than kept static so that
* tests/err_copy_string.c can reach these guards. Both in-library callers
* validate their arguments first, so nothing else can drive them.
*/
akerr_ErrorContext *__akerr_copy_string(char *destination, int capacity,
const char *source)
{
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
PREPARE_ERROR(errctx);
FAIL_NONZERO_RETURN(errctx, (destination == NULL || source == NULL),
AKERR_NULLPOINTER,
"__akerr_copy_string got a NULL %s",
destination == NULL ? "destination buffer" : "source string");
FAIL_NONZERO_RETURN(errctx, (capacity <= 0), AKERR_VALUE,
"__akerr_copy_string got a capacity of %d; a buffer must "
"have room for at least a terminator", capacity);
strncpy(destination, source, (size_t)capacity - 1);
destination[capacity - 1] = '\0';
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
SUCCEED_RETURN(errctx);
}
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
/*
* Compare against each element address rather than testing the address range.
* A range test also accepts pointers into the *interior* of an element, which
* would then be treated as the head of an akerr_ErrorContext and written
* through. Keep this an element-wise scan; it is not a missed optimization.
*/
int akerr_valid_error_address(akerr_ErrorContext *ptr)
{
// Is this within the memory region occupied by AKERR_ARRAY_ERROR?
if ( ptr == NULL ) {
return 1;
}
for ( int i = 0; i < AKERR_MAX_ARRAY_ERROR; i++ ) {
if ( ptr == &AKERR_ARRAY_ERROR[i] ) {
return 1;
}
}
return 0;
}
void akerr_default_logger(const char *fmt, ...)
{
#if defined(AKERR_USE_STDLIB) && AKERR_USE_STDLIB == 1
va_list ap;
va_start(ap, fmt);
vfprintf(stderr, fmt, ap);
va_end(ap);
#else
return;
#endif
}
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
/*
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
* The library naming its own codes.
*
* akerr_init() returns void and runs before any consumer frame exists, so there
* is nothing to PASS an error to: this *is* the top of the stack. FINISH_NORETURN
* is the library's idiom for that position -- the same one main() uses -- so an
* unhandled failure prints its stack trace and goes to
* akerr_handler_unhandled_error, which terminates.
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
*
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
* That is fatal on purpose. The library can only fail to name its own status
* codes if the build is misconfigured -- a name table too small to hold even
* the library's own entries, or a reservation that did not take -- and the
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
* consequence of continuing is every later stack trace in the process printing
* "Unknown Error" for a code the library defines. That is a startup defect, and
* it is far cheaper to see it at init than to debug it from a degraded trace.
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
*
* The generated errno table calls this rather than registering names directly,
* so that all of this control flow lives here in one place.
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
*/
void __akerr_name_library_status(int status, const char *name)
{
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
PREPARE_ERROR(errctx);
ATTEMPT {
CATCH(errctx, akerr_register_status_name(AKERR_LIBRARY_OWNER, status, name));
} CLEANUP {
} PROCESS(errctx) {
} FINISH_NORETURN(errctx);
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
}
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
/*
* Idempotent. `inited` is set before any work so that the registry calls below
* -- and the public registry entry points, which all call akerr_init() so that
* a consumer reserving its range before anything else touches the library
* cannot have that reservation wiped by a later first-use of the pool -- see
* themselves as already initialized instead of recursing.
*/
void akerr_init()
2025-07-20 21:40:14 -04:00
{
static int inited = 0;
if ( inited == 0 ) {
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
inited = 1;
for (int i = 0; i < AKERR_MAX_ARRAY_ERROR; i++ ) {
memset((void *)&AKERR_ARRAY_ERROR[i], 0x00, sizeof(akerr_ErrorContext));
AKERR_ARRAY_ERROR[i].arrayid = i;
AKERR_ARRAY_ERROR[i].stacktracebufptr = (char *)&AKERR_ARRAY_ERROR[i].stacktracebuf;
2025-07-20 21:40:14 -04:00
}
__akerr_last_ignored = NULL;
memset((void *)&__akerr_last_ditch, 0x00, sizeof(akerr_ErrorContext));
__akerr_last_ditch.stacktracebufptr = (char *)&__akerr_last_ditch.stacktracebuf;
if ( akerr_log_method == NULL ) {
akerr_log_method = &akerr_default_logger;
}
akerr_handler_unhandled_error = &akerr_default_handler_unhandled_error;
memset((void *)&akerr_status_names[0], 0x00, sizeof(akerr_status_names));
memset((void *)&akerr_status_ranges[0], 0x00, sizeof(akerr_status_ranges));
akerr_status_name_count = 0;
akerr_status_range_count = 0;
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
/* errno and AKERR_* values are the library-owned compatibility band.
* This must precede every registration below: naming a status is only
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
* permitted inside a reserved range. Terminal for the same reason as
* __akerr_name_library_status(), and handled the same way: without this
* band the library owns nothing, so none of the names below could
* register either. */
PREPARE_ERROR(errctx);
ATTEMPT {
CATCH(errctx, akerr_reserve_status_range(0, AKERR_RESERVED_STATUS_COUNT,
AKERR_LIBRARY_OWNER));
} CLEANUP {
} PROCESS(errctx) {
} FINISH_NORETURN(errctx);
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
/* Every AKERR_* code gets a name; tests/err_error_names.c asserts the
* list is exhaustive so a new code cannot be added without one. */
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
__akerr_name_library_status(AKERR_NULLPOINTER, "Null Pointer Error");
__akerr_name_library_status(AKERR_OUTOFBOUNDS, "Out Of Bounds Error");
__akerr_name_library_status(AKERR_API, "API Error");
__akerr_name_library_status(AKERR_ATTRIBUTE, "Attribute Error");
__akerr_name_library_status(AKERR_TYPE, "Type Error");
__akerr_name_library_status(AKERR_KEY, "Key Error");
__akerr_name_library_status(AKERR_INDEX, "Index Error");
__akerr_name_library_status(AKERR_FORMAT, "Format Error");
__akerr_name_library_status(AKERR_IO, "Input Output Error");
__akerr_name_library_status(AKERR_VALUE, "Value Error");
__akerr_name_library_status(AKERR_RELATIONSHIP, "Relationship Error");
__akerr_name_library_status(AKERR_EOF, "End Of File");
__akerr_name_library_status(AKERR_CIRCULAR_REFERENCE, "Circular Reference Error");
__akerr_name_library_status(AKERR_ITERATOR_BREAK, "Iterator Break");
__akerr_name_library_status(AKERR_NOT_IMPLEMENTED, "Not Implemented");
__akerr_name_library_status(AKERR_BADEXC, "Invalid akerr_ErrorContext");
__akerr_name_library_status(AKERR_STATUS_RANGE_OVERLAP, "Status Range Overlap");
__akerr_name_library_status(AKERR_STATUS_RANGE_FULL, "Status Range Table Full");
__akerr_name_library_status(AKERR_STATUS_RANGE_INVALID, "Invalid Status Range");
__akerr_name_library_status(AKERR_STATUS_NAME_UNRESERVED, "Unreserved Status Name");
__akerr_name_library_status(AKERR_STATUS_NAME_FOREIGN, "Foreign Status Name");
__akerr_name_library_status(AKERR_STATUS_NAME_FULL, "Status Name Registry Full");
__akerr_name_library_status(AKERR_STATUS_NAME_INVALID, "Invalid Status Name");
#if (defined(AKERR_USE_STDLIB) && AKERR_USE_STDLIB == 1) || (!defined(AKERR_USE_STDLIB))
akerr_init_errno();
#endif
2025-07-20 21:40:14 -04:00
}
}
void akerr_default_handler_unhandled_error(akerr_ErrorContext *errctx)
2025-07-20 21:40:14 -04:00
{
if ( errctx == NULL ) {
exit(1);
}
exit(errctx->status);
}
akerr_ErrorContext *akerr_next_error()
2025-07-20 21:40:14 -04:00
{
for (int i = 0; i < AKERR_MAX_ARRAY_ERROR; i++ ) {
if ( AKERR_ARRAY_ERROR[i].refcount == 0 ) {
return &AKERR_ARRAY_ERROR[i];
2025-07-20 21:40:14 -04:00
}
}
return (akerr_ErrorContext *)NULL;
2025-07-20 21:40:14 -04:00
}
akerr_ErrorContext *akerr_release_error(akerr_ErrorContext *err)
2025-07-20 21:40:14 -04:00
{
int oldid = 0;
if ( err == NULL ) {
akerr_ErrorContext *errctx = &__akerr_last_ditch;
FAIL_RETURN(errctx, AKERR_NULLPOINTER, "akerr_release_error got NULL context pointer");
2025-07-20 21:40:14 -04:00
}
if ( err->refcount > 0 ) {
err->refcount -= 1;
}
if ( err->refcount == 0 ) {
oldid = err->arrayid;
memset(err, 0x00, sizeof(akerr_ErrorContext));
2025-07-20 21:40:14 -04:00
err->stacktracebufptr = (char *)&err->stacktracebuf;
err->arrayid = oldid;
2025-07-20 21:40:14 -04:00
return NULL;
}
return err;
}
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
/*
* Scatter the status across the table. Status values are typically dense runs
* (errno 1..N, then a library's block at its base), which linear probing on the
* raw value would pile into one cluster, so mix the bits first.
*/
static unsigned akerr_status_hash(int status)
2025-07-20 21:40:14 -04:00
{
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
unsigned h = (unsigned)status;
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
h ^= h >> 16;
h *= 0x85ebca6bu;
h ^= h >> 13;
h *= 0xc2b2ae35u;
h ^= h >> 16;
return h & (unsigned)(AKERR_STATUS_NAME_SLOTS - 1);
}
/*
* Find the slot holding `status`. With create != 0, claim a free slot for it if
* it is not present yet. Returns NULL when the status is absent and either no
* slot was requested or the registry is full.
*
* The `& (AKERR_STATUS_NAME_SLOTS - 1)` below is load-bearing and fails
* silently: off by one in either direction and the probe indexes past
* akerr_status_names, writing into whatever BSS follows rather than crashing.
* A test that only counts how many names registered before the table filled
* cannot see that -- a probe sequence collapsed to two slots still registers
* "some" names. Any change to the probe sequence, the occupancy cap, or the
* power-of-two assumption needs a test that reads every entry back by its own
* distinct value; tests/err_maxval.c does.
*/
static akerr_StatusName *akerr_status_slot(int status, int create)
{
unsigned slot = akerr_status_hash(status);
for ( int probe = 0; probe < AKERR_STATUS_NAME_SLOTS; probe++ ) {
akerr_StatusName *entry = &akerr_status_names[slot];
if ( entry->used == 0 ) {
if ( create == 0 ||
akerr_status_name_count >= AKERR_MAX_REGISTERED_STATUS_NAMES ) {
return NULL;
}
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
entry->used = 1;
entry->status = status;
entry->name[0] = '\0';
akerr_status_name_count++;
return entry;
}
if ( entry->status == status ) {
return entry;
}
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
slot = (slot + 1u) & (unsigned)(AKERR_STATUS_NAME_SLOTS - 1);
}
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
/* Unreachable: occupancy is capped below the slot count, so the probe above
* always meets a free slot. Present so a future change to that cap cannot
* turn this into a runaway loop. */
return NULL;
}
/* The reservation covering `status`, or NULL if nobody has claimed it. */
static akerr_StatusRange *akerr_range_for_status(int status)
{
for ( int i = 0; i < akerr_status_range_count; i++ ) {
if ( status >= akerr_status_ranges[i].first &&
status <= akerr_status_ranges[i].last ) {
return &akerr_status_ranges[i];
}
}
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
return NULL;
}
/*
* Shared body of both registration entry points. A NULL owner means the caller
* did not identify itself (the legacy two-argument akerr_name_for_status path):
* the status must still lie inside *some* reservation, but we cannot check that
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
* it is the caller's. Every refusal raises an error -- a name that silently
* fails to register degrades into "Unknown Error" in stack traces, which is
* exactly the kind of quiet loss this registry exists to prevent -- so the
* message carries everything a caller needs to see in a stack trace.
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
*/
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
static akerr_ErrorContext AKERR_NOIGNORE *akerr_store_status_name(const char *owner,
int status,
const char *name)
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
{
akerr_StatusRange *range;
akerr_StatusName *entry;
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
PREPARE_ERROR(errctx);
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
FAIL_NONZERO_RETURN(errctx, (name == NULL), AKERR_STATUS_NAME_INVALID,
"Refusing to name status %d for %s: the name is NULL",
status, owner == NULL ? "an unnamed caller" : owner);
FAIL_NONZERO_RETURN(errctx,
(owner != NULL && ( owner[0] == '\0' ||
strlen(owner) >= AKERR_MAX_STATUS_RANGE_OWNER_LENGTH )),
AKERR_STATUS_NAME_INVALID,
"Refusing to name status %d (\"%s\"): the owner string "
"is empty or longer than %d characters",
status, name, AKERR_MAX_STATUS_RANGE_OWNER_LENGTH - 1);
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
range = akerr_range_for_status(status);
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
FAIL_ZERO_RETURN(errctx, range, AKERR_STATUS_NAME_UNRESERVED,
"Refusing to name status %d (\"%s\") for %s: no reserved "
"range contains it. Call akerr_reserve_status_range() first.",
status, name, owner == NULL ? "an unnamed caller" : owner);
FAIL_NONZERO_RETURN(errctx,
(owner != NULL && strcmp(owner, range->owner) != 0),
AKERR_STATUS_NAME_FOREIGN,
"Refusing to name status %d (\"%s\") for %s: that status "
"is in range %d..%d owned by %s.",
status, name, owner,
range->first, range->last, range->owner);
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
entry = akerr_status_slot(status, 1);
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
FAIL_ZERO_RETURN(errctx, entry, AKERR_STATUS_NAME_FULL,
"Status name registry is full (%d entries); dropping name "
"\"%s\" for status %d. Rebuild libakerror with a larger "
"AKERR_STATUS_NAME_SLOTS.",
AKERR_MAX_REGISTERED_STATUS_NAMES, name, status);
PASS(errctx, __akerr_copy_string(entry->name, AKERR_MAX_ERROR_NAME_LENGTH, name));
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
SUCCEED_RETURN(errctx);
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
}
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
/*
* Register a name for a status inside a range the caller reserved. Both strings
* are checked here rather than only inside akerr_store_status_name(): the store
* accepts a NULL owner for the legacy akerr_name_for_status() path, so a NULL
* arriving through *this* entry point would be read as "caller did not identify
* itself" and skip the ownership check entirely.
*/
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
akerr_ErrorContext *akerr_register_status_name(const char *owner, int status, const char *name)
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
{
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
PREPARE_ERROR(errctx);
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
FAIL_NONZERO_RETURN(errctx, (owner == NULL), AKERR_STATUS_NAME_INVALID,
"Refusing to name status %d: the owner string is NULL. "
"Pass the same owner you reserved the range with.",
status);
FAIL_NONZERO_RETURN(errctx, (name == NULL), AKERR_STATUS_NAME_INVALID,
"Refusing to name status %d for %s: the name is NULL",
status, owner);
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
PASS(errctx, akerr_store_status_name(owner, status, name));
SUCCEED_RETURN(errctx);
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
}
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
/*
* Return or set a name. Status magnitude is unrelated to storage size.
*
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
* The set path is the legacy two-argument form. It returns a name, so it cannot
* hand an error back to its caller and cannot raise: it handles the refusal
* here, converting it to the "Unknown Error" sentinel the way any function that
* must return a value converts a caught error into one.
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
* akerr_register_status_name() is the form that raises.
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
*
* The lookup path (name == NULL) deliberately stays clear of all of this. FAIL
* calls it to render a status into a stack trace, so it must not itself need an
* error context.
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
*/
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
char *akerr_name_for_status(int status, char *name)
{
akerr_StatusName *entry;
akerr_init();
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
if ( name != NULL ) {
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
PREPARE_ERROR(errctx);
int refused = 0;
ATTEMPT {
CATCH(errctx, akerr_store_status_name(NULL, status, name));
} CLEANUP {
} PROCESS(errctx) {
} HANDLE_DEFAULT(errctx) {
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
LOG_ERROR_WITH_MESSAGE(errctx, "** REFUSED STATUS NAME **");
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
refused = 1;
} FINISH_NORETURN(errctx);
if ( refused != 0 ) {
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
return "Unknown Error";
}
Enforce status-code ownership and harden the name registry Reservations were advisory bookkeeping: any component could name any status, so the registry only detected declared-range overlap between components that both opted in. Naming a status now requires a reservation. akerr_register_status_name() checks that the range belongs to the caller, and the legacy two-argument akerr_name_for_status() set path, which cannot identify its caller, requires that some reservation covers the status. Every refusal is logged and names the real owner, because a name that fails to register degrades that code to "Unknown Error" in every later stack trace. Fix a reservation made before the first PREPARE_ERROR being silently discarded. akerr_init() clears the tables, so whichever component first triggered it wiped an earlier reservation and the next component to claim the same range was told it was free, producing exactly the undetected aliasing the registry exists to prevent. Every registry entry point now calls akerr_init(), which sets its guard before doing any work so those calls do not recurse. Replace the linear-scan name array with an open-addressed hash table, taking lookup from O(n) to O(1) and raising usable capacity from 512 entries (366 free to consumers after errno registration) to 3072 (~2900 free). Both table sizes are build-time overridable and applied PRIVATE: they live entirely in src/error.c, so raising them cannot desynchronize a library from its consumers the way AKERR_MAX_ERR_VALUE could. Exhausting either table is now logged and returned to the caller rather than silently dropping the entry. No dynamic allocation is introduced; both tables remain file-scope arrays, and the library's undefined-symbol set gains only strcmp and strlen. Register names for AKERR_EOF, AKERR_ITERATOR_BREAK and AKERR_NOT_IMPLEMENTED, which had none and rendered as "Unknown Error" in every stack trace carrying them. err_error_names.c now sweeps the whole AKERR_* offset span so a code added without a name fails there instead of in production traces. Add static assertions that the slot count is a power of two and that AKERR_BADEXC stays inside the library's own 0-255 band, the latter guarding against a host errno space large enough to push library codes into the range consumers are told to allocate from. Set a project version and soname (1.0.0 / libakerror.so.1) so a stale installed library can no longer be silently paired with newer headers, and so akerror.pc ships a real Version field instead of an empty one. Mutation testing surfaced an out-of-bounds probe in the new table that the suite did not catch: masking with SLOTS rather than SLOTS-1 indexes past the array, and err_maxval.c asserted only that some names registered before the table filled, which a collapsed probe sequence still satisfies. It now requires a substantial entry count and reads every entry back by its own distinct name. Tests: 28/28 pass. Coverage 99.4% line / 86.8% branch. Mutation score for src/error.c 74% -> 77.3%. Compatibility: source and ABI break. AKERR_MAX_ERR_VALUE and the __AKERR_ERROR_NAMES data symbol are gone, custom codes must move out of 0-255, and names must be registered against a reserved range. README.md carries the migration steps. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 18:19:25 -04:00
}
entry = akerr_status_slot(status, 0);
if ( entry == NULL ) {
return "Unknown Error";
}
return entry->name;
}
/* Reserve an inclusive status interval and reject collisions. */
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
akerr_ErrorContext *akerr_reserve_status_range(int first_status, int count, const char *owner)
{
int last_status;
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
PREPARE_ERROR(errctx);
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
FAIL_NONZERO_RETURN(errctx,
(count <= 0 || owner == NULL || owner[0] == '\0' ||
strlen(owner) >= AKERR_MAX_STATUS_RANGE_OWNER_LENGTH ||
first_status > INT_MAX - (count - 1)),
AKERR_STATUS_RANGE_INVALID,
"Invalid status range reservation: %d status values from "
"%d for %s (count must be positive, the owner string "
"non-empty and shorter than %d characters, and the range "
"must not overflow int)",
count, first_status, owner == NULL ? "(null)" : owner,
AKERR_MAX_STATUS_RANGE_OWNER_LENGTH);
last_status = first_status + count - 1;
for ( int i = 0; i < akerr_status_range_count; i++ ) {
if ( first_status <= akerr_status_ranges[i].last &&
last_status >= akerr_status_ranges[i].first ) {
if ( first_status == akerr_status_ranges[i].first &&
last_status == akerr_status_ranges[i].last &&
strcmp(owner, akerr_status_ranges[i].owner) == 0 ) {
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
SUCCEED_RETURN(errctx);
}
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
FAIL_RETURN(errctx, AKERR_STATUS_RANGE_OVERLAP,
"Status range %d..%d requested by %s overlaps %d..%d "
"owned by %s",
first_status, last_status, owner,
akerr_status_ranges[i].first,
akerr_status_ranges[i].last,
akerr_status_ranges[i].owner);
}
}
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
FAIL_NONZERO_RETURN(errctx,
(akerr_status_range_count == AKERR_MAX_RESERVED_STATUS_RANGES),
AKERR_STATUS_RANGE_FULL,
"Status range table is full (%d ranges); refusing %d..%d "
"for %s.",
AKERR_MAX_RESERVED_STATUS_RANGES,
first_status, last_status, owner);
/* The owner copy is what commits the entry, so the count only advances
* once it has succeeded -- a half-written reservation would claim the
* range under an empty owner nobody could ever match. */
akerr_status_ranges[akerr_status_range_count].first = first_status;
akerr_status_ranges[akerr_status_range_count].last = last_status;
Use the library's own error idioms inside the library Four things in src/error.c did by hand what the macros already do, or skipped checks the library would have caught for a consumer. akerr_copy_string() returned void and validated only its capacity, while writing through a caller-supplied pointer for a caller-supplied length. It is now __akerr_copy_string() and raises: AKERR_NULLPOINTER for a NULL destination or source, AKERR_VALUE for a capacity with no room for a terminator. Both call sites PASS it, and the owner copy in akerr_reserve_status_range() now gates the commit, so a failed copy cannot leave a range claimed under an empty owner. It is exported under the internal prefix rather than static so tests/err_copy_string.c can drive those guards; nothing else can reach them. __akerr_name_library_status() and the band reservation in akerr_init() hand-rolled the log/handler/release sequence. Both now use ATTEMPT/CATCH/PROCESS/FINISH_NORETURN. PASS does not fit: both sites are void and have no caller to propagate to, so the terminal form of the same idiom is the right one -- an unhandled failure prints its stack trace and goes to akerr_handler_unhandled_error, which terminates, exactly as before but without the bespoke plumbing. The legacy set path in akerr_name_for_status() had the same shape and now handles its refusal with HANDLE_DEFAULT, converting it to the "Unknown Error" sentinel. Every remaining `if (x) { FAIL_RETURN }` in the registry is now FAIL_ZERO_RETURN or FAIL_NONZERO_RETURN, and akerr_register_status_name() checks both owner and name before passing either down -- akerr_store_status_name() reads a NULL owner as "caller did not identify itself" for the legacy path, so a NULL arriving through the owned entry point would have skipped the ownership check entirely. New tests: err_copy_string (the guards above), err_library_status_fatal (WILL_FAIL -- proves a refused library-status registration terminates). Tests: ctest 31/31, mutation 80.7% (was 77.5%), line coverage 98.9%. Branch coverage on src/error.c drops 64.5% -> 50.4%, just over its gate: each FAIL_* site carries ~6 branch outcomes of error-construction machinery that only run when that failure fires, and each PASS around a call that cannot fail carries ~25, so added validation lowers the ratio by construction. Recorded in TODO.md item 7. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 21:53:33 -04:00
PASS(errctx, __akerr_copy_string(akerr_status_ranges[akerr_status_range_count].owner,
AKERR_MAX_STATUS_RANGE_OWNER_LENGTH, owner));
akerr_status_range_count++;
Raise errors from the status registry instead of returning codes akerr_reserve_status_range() and akerr_register_status_name() returned private int enumerations, which was the one place in the library where a failure was not an akerr_ErrorContext *. They now return one like everything else: NULL on success, and on refusal an error whose status is a real code in the library's reserved band, so it can be CATCH-ed, HANDLE-d, PASS-ed, or left to propagate into a stack trace. Both are marked AKERR_NOIGNORE, so discarding the result warns at compile time. AKERR_STATUS_RANGE_OK and AKERR_STATUS_NAME_OK are gone; the remaining seven codes move into the AKERR_* offset span and get registered names. AKERR_LAST_LIBRARY_STATUS replaces AKERR_BADEXC as the top of that span in the reserved-band static assert and the exhaustiveness sweep. The refusal detail that used to go straight to akerr_log_method now travels in the error message, so a caller that handles the error decides whether it is reported. The two-argument akerr_name_for_status() set path is the exception: it returns a name and cannot raise, so it logs and releases. akerr_init() likewise has no caller to raise into, so failing to reserve its own band or name its own codes is logged and fatal -- that can only happen on a misconfigured build, and continuing would degrade every later stack trace to "Unknown Error". Move the 1.0.0 upgrade notice out of README.md into UPGRADING.md and rewrite its return-code tables in terms of the statuses now raised. Tests: ctest 29/29, coverage 97.5% line / 64.5% branch, mutation 77.5% (was 77.6%; the new survivors are the fatal init path, which needs a library built with an undersized name table -- TODO item 7). Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-30 20:58:47 -04:00
SUCCEED_RETURN(errctx);
2025-07-20 21:40:14 -04:00
}