andrew/libakerror
1
0
Fork 0
Code Issues 10 Pull Requests Actions 1 Packages Projects Releases Wiki Activity

Release ignored error contexts #21

Merged
andrew merged 3 commits from 14 into main 2026-08-04 11:28:30 -04:00
Conversation 5 Commits 3 Files Changed 9 +58 -36
logikoma commented 2026-08-03 12:57:11 -04:00
Collaborator
Copy Link

Closes #14.

Summary

  • release the error context after IGNORE() logs it
  • retain a per-thread static snapshot of the ignored error for debugging
  • repair the copied stack-trace cursor so the snapshot remains self-contained
  • exercise more ignored failures than the pool has slots
  • update threaded coverage and lifecycle documentation

Tests

  • cmake -S . -B build
  • cmake --build build
  • ctest --test-dir build --output-on-failure (36/36)
  • scripts/thread_test.sh (36/36 under ThreadSanitizer)
  • cmake -S . -B build/none -DAKERR_THREADS=none
  • ctest --test-dir build/none --output-on-failure (32/32)
Closes #14. ## Summary - release the error context after IGNORE() logs it - retain a per-thread static snapshot of the ignored error for debugging - repair the copied stack-trace cursor so the snapshot remains self-contained - exercise more ignored failures than the pool has slots - update threaded coverage and lifecycle documentation ## Tests - cmake -S . -B build - cmake --build build - ctest --test-dir build --output-on-failure (36/36) - scripts/thread_test.sh (36/36 under ThreadSanitizer) - cmake -S . -B build/none -DAKERR_THREADS=none - ctest --test-dir build/none --output-on-failure (32/32)
logikoma added 1 commit 2026-08-03 12:57:11 -04:00
Release ignored error contexts
All checks were successful
libakerror CI Build / cmake_build (push) Successful in 2m52s
Details
libakerror CI Build / coverage (push) Successful in 2m48s
Details
libakerror CI Build / mutation_test (push) Successful in 44m14s
Details
eec5726d91
logikoma force-pushed 14 from eec5726d91 to 4fe7571329 2026-08-03 13:46:44 -04:00 Compare
andrew reviewed 2026-08-03 13:59:34 -04:00
andrew reviewed 2026-08-03 14:01:28 -04:00
tests/err_threads_pool.c Outdated
@@ -105,2 +100,3 @@
/* IGNORE's scratch pointer is thread-local while logging and cleared after
* release. Concurrent ignored errors must all return their pool slots. */
IGNORE(ignorable(arg));
AKERR_TCHECK(arg, __akerr_last_ignored != NULL);
andrew commented 2026-08-03 14:01:28 -04:00
Owner
Copy Link

IGNORE() retained a reference to the last ignored error on purpose, so that subsequent errors that may be related could reference it. Instead of completely abandoning it (and the pattern that allows it), why not turn __akerr_last_ignored into a static variable, IGNORE() does a memcpy() into it from the exception being ignored, then releases the exception? That seems like it would give us the best of both worlds.

IGNORE() retained a reference to the last ignored error on purpose, so that subsequent errors that may be related could reference it. Instead of completely abandoning it (and the pattern that allows it), why not turn __akerr_last_ignored into a static variable, IGNORE() does a `memcpy()` into it from the exception being ignored, then releases the exception? That seems like it would give us the best of both worlds.
andrew commented 2026-08-03 18:27:52 -04:00
Owner
Copy Link

IGNORE() retained a reference to the last ignored error on purpose, so that subsequent errors that may be related could reference it. Instead of completely abandoning it (and the pattern that allows it), why not turn __akerr_last_ignored into a static variable, IGNORE() does a memcpy() into it from the exception being ignored, then releases the exception? That seems like it would give us the best of both worlds.

@logikoma ^ implement this

> IGNORE() retained a reference to the last ignored error on purpose, so that subsequent errors that may be related could reference it. Instead of completely abandoning it (and the pattern that allows it), why not turn __akerr_last_ignored into a static variable, IGNORE() does a `memcpy()` into it from the exception being ignored, then releases the exception? That seems like it would give us the best of both worlds. @logikoma ^ implement this
andrew marked this conversation as resolved
logikoma requested review from andrew 2026-08-03 18:22:29 -04:00
logikoma added 1 commit 2026-08-04 10:45:52 -04:00
Preserve ignored error snapshots
Some checks failed
libakerror CI Build / coverage (push) Successful in 3m39s
Details
libakerror CI Build / cmake_build (push) Successful in 6m3s
Details
libakerror CI Build / mutation_test (push) Has been cancelled
Details
55090d2419
logikoma commented 2026-08-04 10:46:02 -04:00
Author
Collaborator
Copy Link

Implemented the requested review change. IGNORE() now copies the error into a per-thread static snapshot with a repaired stack-trace cursor, logs that snapshot, and releases the original pool context. Tests verify the snapshot survives slot reuse in both single-threaded and threaded runs. Full CTest, ThreadSanitizer, and no-thread CTest pass.

Implemented the requested review change. IGNORE() now copies the error into a per-thread static snapshot with a repaired stack-trace cursor, logs that snapshot, and releases the original pool context. Tests verify the snapshot survives slot reuse in both single-threaded and threaded runs. Full CTest, ThreadSanitizer, and no-thread CTest pass.
andrew reviewed 2026-08-04 10:58:14 -04:00
include/akerror.tmpl.h Outdated
@@ -441,0 +438,4 @@
do { \
akerr_ErrorContext *__akerr_ignored = __stmt; \
if ( __akerr_ignored != NULL ) { \
memcpy(&__akerr_last_ignored, __akerr_ignored, \
andrew commented 2026-08-04 10:58:14 -04:00
Owner
Copy Link

@logikoma the __akerr_last_ignored variable is only really useful to users of the public facing API, so I'm not sure it makes sense to prefix it with __ as part of the private API. Let's remove the __ prefix on this variable and make it public.

@logikoma the `__akerr_last_ignored` variable is only really useful to users of the public facing API, so I'm not sure it makes sense to prefix it with `__` as part of the private API. Let's remove the `__` prefix on this variable and make it public.
logikoma marked this conversation as resolved
logikoma added 1 commit 2026-08-04 11:25:31 -04:00
Expose ignored error snapshot
All checks were successful
libakerror CI Build / coverage (push) Successful in 3m18s
Details
libakerror CI Build / cmake_build (push) Successful in 10m11s
Details
libakerror CI Build / mutation_test (push) Successful in 45m33s
Details
5a269ea01b 
Co-Authored-By: Andrew Kesterson <andrew@aklabs.net>
logikoma commented 2026-08-04 11:25:50 -04:00
Author
Collaborator
Copy Link

Implemented in 5a269ea. IGNORE() now copies the ignored context into the per-thread static snapshot, repairs the copied stack-trace cursor, logs it, and releases the pool context. The public snapshot is now named akerr_last_ignored. Normal tests pass 36/36; AKERR_THREADS=none passes 32/32.

Implemented in 5a269ea. IGNORE() now copies the ignored context into the per-thread static snapshot, repairs the copied stack-trace cursor, logs it, and releases the pool context. The public snapshot is now named akerr_last_ignored. Normal tests pass 36/36; AKERR_THREADS=none passes 32/32.
andrew merged commit c82f5fe695 into main 2026-08-04 11:28:30 -04:00
Sign in to join this conversation.
Reviewers
No Reviewers
andrew
Labels
Clear labels
abi-break
Cannot land without a soname bump
 api-gap
A caller cannot express something
 blast-radius:high
Ordering: highest blast radius
 blast-radius:low
Ordering: lowest blast radius
 blast-radius:medium
Ordering: middling blast radius
 defect
Wrong behaviour reachable from the public API
 design-decision
Needs a decision written down before code
 docs
Documentation, comments, AGENTS.md
 hygiene
Style, dead code, consistency
 packaging
Build, install, pkg-config, CMake config
 performance
Has a measurement behind it
 status::blocked
Waiting on something else -- a decision, a dependency, or another issue
 status::grooming
Not yet worked out: scope, approach or priority still being settled
 status::ready
Understood well enough to start: scope and approach are settled
 test-coverage
A gap in the suites, not a defect
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
logikoma tachikoma andrew
No Assignees
2 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: andrew/libakerror#21
Delete Branch "14"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?