fix: preserve native HTTPS startup failure diagnostics - #130
Merged
Merged
Conversation
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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 detached native HTTPS helper can fail before publishing its endpoint. The acquisition wrapper discards that failure, and a later cleanup error can replace it with an unhelpful missing-file error. Preserve a bounded startup classification through both boundaries so the next failed run identifies where startup stopped.
The helper writes a private diagnostic pinned to its exact owner generation and configuration digest. Only fixed stages and seven reviewed native error codes are admitted; no child stderr, paths, application values or arbitrary error messages cross the boundary. Acquisition diagnostics survive unconfirmed graph, HTTPS, temporary-file, lifecycle and finalization cleanup. A missing diagnostic remains unknown. The record never acknowledges retirement, authorizes replacement/adoption, changes confirmation ordering or retries an operation.
Validation: focused owner/socket, HTTPS and project-start regressions cover detached failure, hard exit without a record, foreign/stale/malformed/symlink records, and every later cleanup masking path. Root typecheck/lint, changed-test lint, staged privacy and whitespace checks pass. Final full CLI gate passes: 1833 tests, 67 skips, 0 failures across 258 files (128.79s). An earlier run had one timeout in unchanged dependency-startup-progress while waiting for its fake Docker subprocess marker; that file passed 6/6 in isolation and the unchanged full rerun passed. The failed-run evidence is retained; no test assertion or timeout was changed. DB gates are cached; CLI gates execute.
Release signal:
fix, no migration. This corrects diagnosis, not the unproven underlying helper-startup cause. Fresh native qualification remains open; the retained failing fixture is preserved and no native runtime was started by this change.