fix(artifact): ARTIFACT task & embedded-extraction reliability + verification harness#2272
Draft
pirhoo wants to merge 19 commits into
Draft
fix(artifact): ARTIFACT task & embedded-extraction reliability + verification harness#2272pirhoo wants to merge 19 commits into
pirhoo wants to merge 19 commits into
Conversation
pirhoo
force-pushed
the
audit/artifact-embedded-reliability
branch
from
July 16, 2026 16:20
1fe2e54 to
afe08cf
Compare
… corrupt members killed_run_then_rerun_converges_to_full_coverage and corrupt_member_is_reported_visibly_and_siblings_are_covered pass; parallelism_4_produces_identical_coverage is committed RED as an F1 (embedded-extraction race) audit finding, per the branch's chaos-test policy.
executor.shutdownNow() only requests shutdown; it does not wait for worker threads to unwind. On the cancellation path, cancel() interrupts the task thread first, so future.get() throws InterruptedException and reaches the finally block before workers have actually stopped. That left a race where drainResidualPoison() could run concurrently with a worker mid-relay of STRING_POISON (dequeued, about to offer() it back), letting a stale poison slip past the drain and strand the next run. Wait for the pool to terminate (bounded, 30s) before draining, mirroring the awaitTermination precedent in IndexTask. The wait's own InterruptedException is caught locally and re-interrupts the thread so the original InterruptedException from future.get() still propagates unmasked.
…eds stay retrievable
…age holes A terminal manifest entry + a readable (even 0-byte) source is a correct representation of an empty embed on real OSTs (empty mail-item nodes, empty message bodies), not a loss — extract-lib's atomic writes preclude truncation-to-zero for our own writer. Redefine COVERED as terminal manifest + getSource not throwing; track 0-byte-but-present docs in a new Report.empties() tally surfaced as an FYI line in summary(), instead of flagging them as holes. Report.complete() and ArtifactCoverageMain's exit code stay gated on holes only, so empties-only runs still exit 0.
pirhoo
force-pushed
the
audit/artifact-embedded-reliability
branch
from
July 16, 2026 17:14
afe08cf to
528a769
Compare
…workers don't exit early
…defects are diagnosable
… at a plain re-run
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.
Makes the ARTIFACT task reliable for embedded-document retrieval and adds a verification harness. Paired with extract-lib 9.15.5, a real 66 MB OST now yields 289/289 embeds retrievable (up from 16). Do not merge before extract-lib 9.15.5 is published to Maven.
ArtifactCoverageChecker+ArtifactCoverageMainto audit manifest + source completeness of any projectNastyCorpusfixtures, end-to-end coverage test, and chaos tests (kill/resume, parallelism, corrupt member)