Repository navigation
Conversation
Not up to standards ⛔
|
vt128
force-pushed
the
codex/slt-22-paired-map-entries-20261107
branch
from
November 7, 2026 04:47
50a1a46 to
5534c34
Compare
vt128
force-pushed
the
codex/fix-collection-pointers-20261107
branch
from
November 7, 2026 08:45
2da39ac to
7076099
Compare
vt128
changed the base branch from
codex/fix-collection-pointers-20261107
to
master
November 7, 2026 08:46
vt128
changed the base branch from
master
to
codex/fix-collection-pointers-20261107
November 7, 2026 09:37
vt128
force-pushed
the
codex/slt-22-paired-map-entries-20261107
branch
from
November 7, 2026 09:39
5534c34 to
7d93576
Compare
vt128
force-pushed
the
codex/fix-collection-pointers-20261107
branch
from
November 7, 2026 15:03
7076099 to
0d309e9
Compare
vt128
force-pushed
the
codex/slt-22-paired-map-entries-20261107
branch
from
November 7, 2026 15:03
7d93576 to
da33986
Compare
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## codex/fix-collection-pointers-20261107 #87 +/- ##
==========================================================================
+ Coverage 93.19% 93.48% +0.29%
==========================================================================
Files 8 8
Lines 1734 1751 +17
==========================================================================
+ Hits 1616 1637 +21
+ Misses 117 113 -4
Partials 1 1 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
vt128
force-pushed
the
codex/slt-22-paired-map-entries-20261107
branch
from
November 9, 2026 15:08
da33986 to
711421f
Compare
vt128
force-pushed
the
codex/fix-collection-pointers-20261107
branch
2 times, most recently
from
November 10, 2026 03:27
500a12c to
b7c0f45
Compare
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.
NaN-keyed Go map entries could lose their original values during materialization, and a NaN key could hide a reference cycle from String safety checks. Capture key/value pairs with MapRange, keep deterministic ordering, and use the paired values for conversion and cycle detection. When distinct Go NaN keys collide under the fixed Starlark interpreter's equality rules, MakeDict returns an error instead of dropping data. Get/Delete equality semantics are unchanged.
Current branch maintenance: rebased onto the documentation changes on master
cd5f885d(with the existing parent stack). All 25 comment-free Go ASTs/compiler directives, 17 test files and module files match the prior candidate. Fresh Go 1.27.3 / Go 1.19.13 build, race tests twice and vet pass; the normal commit hook passes. The old Go formatter has one pre-existing test-layout difference, retained without changing tests. Prior coverage/downstream/performance receipts retain their original heads. This maintenance does not resolve the performance HOLD; keep Draft and require fresh cloud gates.Regression tests stay in the existing thematic files. They cover scalar/composite NaN keys, original paired values, script items/values, repeated-NaN conversion errors, self-reference and bounded-child/parent coverage. This PR stacks on the optimized #86; its runtime/module files and 231 regular test/example/fuzz function bodies match the verified combined candidate. Forty persistent benchmarks remain inherited from #86.
Retained historical runtime evidence (original source labels preserved): the combined candidate passes host Go 1.27.3 and offline Go 1.19.13 build/vet/full race tests twice. Relative to the verified #86 tree, total statement coverage rises from 1296/1398 (92.703863%) to 1312/1411 (92.983700%) on both compilers. All changed-production-line-intersecting blocks are hit: host 19 blocks/32 statements, floor 20 blocks/36 statements. Its 80 active downstream legs plus four archived diagnostic legs pass with identical paired package coverage. These are statement/block measurements, not branch-edge coverage.
Performance gate: HOLD together with #86. The L1 host40/floor14 set has no Holm-significant time regression, but integrated Starlet Call/Print1x4096/CustomHandler retains a positive timing signal: initial paired +5.6897%, Holm .0498; longer same-binary 12-pair × 1-second confirmation paired +3.704%, 95% paired-bootstrap interval [+0.278%, +9.203%], Holm .168, with unchanged target B/op/allocations. The longer result does not establish equivalence. Raw results and the adjacent-control outlier are retained; no unproven claim of zero performance loss is made.
Refs SLT-22. Keep Draft until the performance gate is resolved. Some required checks may be withheld for a stacked base branch; absent checks are not passes. This rebased head needs fresh exact-head cloud gates; old-head green checks are not reused. No floor, interpreter pin or module change. No tag or release.
Superseded by combined collection-safety PR #94, merged as a7a2ede after all required checks, scoped safety-cost acceptance and downstream verification. This historical PR is closed without merging because its complete change is delivered by #94. Original source heads, performance measurements and earlier validation failures remain preserved.