Repair macOS sweep preparation, storage, and completion reporting - #21
Merged
Merged
Conversation
added 4 commits
September 22, 2026 13:10
Scope retention to the active campaign while guarding physical free space, persist and honor resume allowances, and report failure or interruption truthfully. Reuse frozen-source preparation per clip, continue real checkpoints, and retire only verified warmup or server-accepted output bytes. Constraint: Preserve protocol 7.1 recipes, full repetitions, and rc.4 Windows fixes Rejected: Raise only the custom launcher's limit | downloaded clients would still fail Confidence: high Scope-risk: moderate Directive: Count owned originals plus publication staging; unrelated journals are independent Tested: 487 client tests and 64 subtests, including canonical archive verification; 88 focused storage, resume, routing and publication checks Not-tested: Packaged Small validation and public distribution verification pending
Expose rc.5 with the verified new macOS disk-image digest and preserve the verified rc.4 Windows and original Linux assets. Keep the previous release's actual checksums and describe the platform-specific provenance accurately. Constraint: Frontend-only publication; preserve the live API, database and capacity configuration Confidence: high Scope-risk: narrow Tested: 78 frontend tests; eslint and TypeScript checks; all four release-asset SHA-256 digests Not-tested: Public rollout and anonymous redownload pending
The guided Continue action bypassed sweep sizing and checkpoint continuation, so a saved large plan inherited the single-run 100-attempt cap and stopped at its first hourly checkpoint. Restore the plan-sized cap and continue the exact saved tasks without changing recipes or repetition order. Older large journals also inherit the same default retention allowance as new large sweeps. Constraint: Explicit attempt and duration limits remain authoritative Confidence: high Scope-risk: narrow Tested: 83 resume, routing, checkpoint and durability tests; parser-based continuation regressions
Stamp the final clean b0607f0 macOS package after verifying its mounted executable and signature. Windows and Linux asset identities remain unchanged. Confidence: high Scope-risk: narrow Tested: Mounted DMG identity/signature; 12 download-page and release-config tests
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.
A downloaded macOS client could spend hours repeatedly probing the same seven clips, reject a Small sweep because another saved campaign exceeded its storage allowance, and then display “complete” after returning an error. This repairs preparation, separates campaign retention from shared publication storage, preserves explicit and saved limits on resume, and reports failure/interruption truthfully. Checkpoints upload completed groups and continue without trimming repetitions; only verified warmups and server-accepted artifacts are retired.
The rc.5 download metadata points to a new macOS DMG built from clean source
b0607f0434f14d5d699b06039d01ed41a2ab68fb. The Windows binaries are unchanged verified rc.4 files and Linux retains its accepted original build. Protocol, frozen clips, runtime lock, and scientific settings remain unchanged. Publication will update only the frontend download surface, preserving production API/database/capacity settings.Validation: 487 client tests and 64 subtests passed (the disk-heavy canonical archive test was run separately); 78 frontend tests, eslint, and TypeScript checks passed. All four release-asset hashes were verified, and the macOS DMG passed a fresh read-only mount and code-signature/content audit. The final mounted macOS package completed its normal guided Small sweep against the real default queue retaining over 2 GB from a Large campaign: 20 genuine encodes, 14 submitted measurements, zero queued/skipped/failed, exit 0, in 3m 25s. Saved-sweep resume additionally restores the full plan attempt cap and continues checkpoints unless the operator sets an explicit duration; parser-based regressions cover both modes. Final DMG SHA-256:
cc6c6503293c8f01b223e1fac9a5da56888352d39c5e777cf90d23acec8aa8c6.