Repository navigation
Share inline timestamp decoding across qualified converters - #7
Merged
Merged
Conversation
Keep the uid probe and TLS daemon on the same immutable image. Refs Huly QB-1107.
Re-read the referenced implementation and correct stale coordinates or wording. Preserve executable bytes outside comments and physical line counts; no runtime or benchmark fixture changes. Huly: QB-1102
Delegate qualified binary conversions to the canonical wall_time specialization so buffer checks cannot drift. Cover ground truth, accepted buffer shapes and failure diagnostics.
|
Connected to Huly®: QB-1119 |
mbelhadi
marked this pull request as ready for review
October 23, 2026 12:09
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Share one explicitly inline template decoder between canonical and qualified PostgreSQL timestamp conversions, retaining the canonical non-template entry point in the .cpp file. This removes the qualified converter's extra out-of-line delegation while preserving eight-byte and legacy-prefixed input, exact malformed-length diagnostics, byte order, negative fractional timestamps and cv/ref parity. The cumulative PR also pins the PostgreSQL CI service image by digest and updates primitive-reader references.
The final composed Windows Release and Linux Release, ASan/UBSan, TSan and coverage suites passed with zero executed-test failures and zero build warnings; platform/configuration/service skips and disabled cases were reviewed explicitly. Eight affected URI/HTTP/PG owners also passed 20 shuffled repetitions on each OS (5340 successes, no skips per OS). Fresh coverage of the final changed paths was reviewed. Runtime sources remain byte-identical to the validation seal; the root PR records the separately validated two-fixture CI follow-up.
The five affected PostgreSQL owners passed 122 cases times 20 shuffled repetitions per OS. Fresh coverage observed both outcomes of the decoder's source conditions for the canonical and three qualified instantiations; this is not proof over every value or library exception arc.
Positive means slower. Const/ConstRef intervals cross zero on both OSes; that does not prove equivalence. Linux canonical decoding retains +0.774% [0.254, 0.981], not a demonstrated unavoidable cost. The eight-block targeted protocol uses new Windows real-time measurements and the existing Linux CPU-time campaign with final executable/dependency identity. No epochs are pooled, and these three cases do not replace the historical actor or whole-network screens.
Documentation and static guards pass on the final source. The original full controls run remains recorded as FAILED: 12 namespace-control misses were caused by the URI preprocessor brace shape. After a repair preserving compiled tokens and line count, the entire namespace battery passed (11 caught, 5 confirmed, 0 missed, 0 skipped), and ordinary verify passed with 68/68 Mermaid renders. Combining that supplement with the 36 retained, unchanged batteries gives 595 caught / 362 confirmed / 0 missed / 4 structural Windows skips. This is a documented composite result, not a new full controls run.
Keep this PR in draft while the coordinated repository review and dependent CI are completed. Module CI must use the matching published qb ref (the workflows expose qb_ref) or the pinned superproject composition; automatic PR checks against qb/develop do not validate a mixed ABI. Publishing these branches and draft PRs does not authorize merging develop.
Evidence and outstanding decisions are tracked in QB-985; the selected ActorId implementation and its accepted trade-offs are tracked in QB-970.
The complete composition, validation table and dependency order are in qb-dev #10.