fix(file_io): write_json atomically (temp + os.replace) to survive interrupted writes#96
Open
gadievron wants to merge 1 commit into
Open
Conversation
…terrupted writes
write_json opened the target in "w" (truncating it to 0 bytes) and then streamed json.dump.
A crash mid-serialization (SIGKILL / OOM / power loss) left the target partial or empty and
destroyed the prior good copy, so the next read_json raised JSONDecodeError -- e.g. a
corrupted checkpoint.json that breaks resume.
Serialize to a temp file in the same directory, flush + fsync, then os.replace onto the
target (atomic rename). An interrupted write leaves only the temp file (unlinked on a caught
error); the existing target is never truncated. All write_json call sites inherit the fix.
The temp uses a `.tmp` suffix (not `.json`) so a hard-crash leftover can't be miscounted by
directory scanners that do os.listdir + endswith(".json") (e.g. core/checkpoint.py).
Tests: tests/test_file_io_atomic_write.py (2 cases: interrupted write preserves the prior
file + no temp leak; normal round-trip incl. non-ASCII). RED 1 failed (JSONDecodeError) ->
GREEN 2 passed (venv py3.11). ruff clean.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
gadievron
requested review from
ar7casper,
dgeyshis,
shahar-davidson,
sounil and
yotamleo
as code owners
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
write_json opened the target in "w" (truncating it to 0 bytes) and then streamed json.dump.
A crash mid-serialization (SIGKILL / OOM / power loss) left the target partial or empty and
destroyed the prior good copy, so the next read_json raised JSONDecodeError -- e.g. a
corrupted checkpoint.json that breaks resume.
Serialize to a temp file in the same directory, flush + fsync, then os.replace onto the
target (atomic rename). An interrupted write leaves only the temp file (unlinked on a caught
error); the existing target is never truncated. All write_json call sites inherit the fix.
The temp uses a
.tmpsuffix (not.json) so a hard-crash leftover can't be miscounted bydirectory scanners that do os.listdir + endswith(".json") (e.g. core/checkpoint.py).
Tests: tests/test_file_io_atomic_write.py (2 cases: interrupted write preserves the prior
file + no temp leak; normal round-trip incl. non-ASCII). RED 1 failed (JSONDecodeError) ->
GREEN 2 passed (venv py3.11). ruff clean.
Co-Authored-By: Claude Opus 4.7 (1M context) noreply@anthropic.com