You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
[fix] make collection conversion and NaN mutation safe - #94
Address three collection-conversion correctness problems with the existing small fixes:
Preserve accepted nil or indirect pointer elements as opaque Go values in no-error APIs, and stop cyclic pointer/interface unwrapping with local cycle state. Static unsupported types, host identity and method precedence retain their contracts.
Snapshot map keys and values together so NaN-like keys retain their values and cycle checks inspect the actual entries. MakeDict reports NaN representation conflicts instead of silently merging values.
Keep host-map aliases intact: on Go 1.19, Clear rejects non-reflexive keys before changing anything, and popitem reports a failed deletion instead of claiming success.
This combines the unchanged SLT-21/22/23 patches on master 59694c47; no new API, dependency or interpreter/floor change. Eight fresh host/Go 1.19 baseline-and-prefix checks pass, including race tests twice and vet. Statement coverage increases from 1276/1385 to 1320/1418 on both compilers, with no old covered hits lost and all new or rewritten statements covered. Existing tests and error contracts are retained.
Known costs remain disclosed: prior combined integration measurements retained a roughly 3.7–5.7% paired timing signal in one Call/4 KiB-print path, with cause unconfirmed and 4192 B/op / 5 allocs/op unchanged. Selected Go 1.19 Clear fixtures cost about 7–13% more time without additional allocations. Tested paired-entry Items/MakeDict/map-argument paths retain or reduce bytes and remove one allocation. These necessary safety costs are accepted under the safety-first, simple-implementation criterion; no new sampling, performance equivalence or universal zero-overhead claim is made.
Fresh full downstream verification passes: 168 configurations across the baseline and all three patch prefixes, with 126 exact profile comparisons and no lost coverage. Published dependency pins are retained; Box eval skips on its older Starlet pin are explicitly classified. A separate 20-configuration joint check with current Starlet b8e0c8f2 passes, including actual Box eval tests, plus four fresh CLI binaries/eight public examples. Source and embedded dependency metadata are bound to each cohort.
The CI-only successor 852175b keeps the complete tested git tree unchanged. After Actions was re-enabled, its exact-head Build, default CodeQL and all 14 required checks, including both Codecov contexts, succeeded. The latest delivered CLI source c2a00525 also passes eight paired main/examples configurations on production Go 1.27.4 and minimum Go 1.26.0: build, full race tests twice, vet and eight real hello/qrcard executions. Both compiler-specific CLI profiles remain 797/881 covered with no previously covered blocks lost. Default Starlet v0.2.3 and all other published pins stay fixed; only Starlight is replaced in the candidate consumer fixtures.
Historical local cohorts retain their original source labels; no benchmark is resampled. Earlier PRs #86–#88 are included by this topic and will be closed as superseded after it actually merges. Normal review/watch/fresh-head gates remain mandatory; no tags or releases are created.
Coverage variation is the difference between the coverage for the head and common ancestor commits of the pull request branch: <coverage of head commit> - <coverage of common ancestor commit>
Diff coverage is the percentage of lines that are covered by tests out of the coverable lines that the pull request added or modified: <covered lines added or modified>/<coverable lines added or modified> * 100%
NEW Get contextual insights on your PRs based on Codacy's metrics, along with PR and Jira context, without leaving GitHub. Enable AI reviewer TIP This summary will be updated as you push new changes.
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
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.
Address three collection-conversion correctness problems with the existing small fixes:
MakeDictreports NaN representation conflicts instead of silently merging values.Clearrejects non-reflexive keys before changing anything, andpopitemreports a failed deletion instead of claiming success.This combines the unchanged SLT-21/22/23 patches on master
59694c47; no new API, dependency or interpreter/floor change. Eight fresh host/Go 1.19 baseline-and-prefix checks pass, including race tests twice and vet. Statement coverage increases from 1276/1385 to 1320/1418 on both compilers, with no old covered hits lost and all new or rewritten statements covered. Existing tests and error contracts are retained.Known costs remain disclosed: prior combined integration measurements retained a roughly 3.7–5.7% paired timing signal in one Call/4 KiB-print path, with cause unconfirmed and 4192 B/op / 5 allocs/op unchanged. Selected Go 1.19
Clearfixtures cost about 7–13% more time without additional allocations. Tested paired-entry Items/MakeDict/map-argument paths retain or reduce bytes and remove one allocation. These necessary safety costs are accepted under the safety-first, simple-implementation criterion; no new sampling, performance equivalence or universal zero-overhead claim is made.Fresh full downstream verification passes: 168 configurations across the baseline and all three patch prefixes, with 126 exact profile comparisons and no lost coverage. Published dependency pins are retained; Box eval skips on its older Starlet pin are explicitly classified. A separate 20-configuration joint check with current Starlet b8e0c8f2 passes, including actual Box eval tests, plus four fresh CLI binaries/eight public examples. Source and embedded dependency metadata are bound to each cohort.
The CI-only successor
852175bkeeps the complete tested git tree unchanged. After Actions was re-enabled, its exact-head Build, default CodeQL and all 14 required checks, including both Codecov contexts, succeeded. The latest delivered CLI sourcec2a00525also passes eight paired main/examples configurations on production Go 1.27.4 and minimum Go 1.26.0: build, full race tests twice, vet and eight real hello/qrcard executions. Both compiler-specific CLI profiles remain 797/881 covered with no previously covered blocks lost. Default Starlet v0.2.3 and all other published pins stay fixed; only Starlight is replaced in the candidate consumer fixtures.Historical local cohorts retain their original source labels; no benchmark is resampled. Earlier PRs #86–#88 are included by this topic and will be closed as superseded after it actually merges. Normal review/watch/fresh-head gates remain mandatory; no tags or releases are created.