chore(deps): Bump github.com/odvcencio/gotreesitter from 0.20.5 to 0.47.0#62
Merged
townsendmerino merged 1 commit intoJul 24, 2026
Conversation
dependabot
Bot
force-pushed
the
dependabot/go_modules/github.com/odvcencio/gotreesitter-0.47.0
branch
from
July 24, 2026 14:08
e1fc69d to
446476a
Compare
Owner
|
@dependabot rebase |
Bumps [github.com/odvcencio/gotreesitter](https://github.com/odvcencio/gotreesitter) from 0.20.5 to 0.47.0. - [Release notes](https://github.com/odvcencio/gotreesitter/releases) - [Changelog](https://github.com/odvcencio/gotreesitter/blob/main/CHANGELOG.md) - [Commits](odvcencio/gotreesitter@v0.20.5...v0.47.0) --- updated-dependencies: - dependency-name: github.com/odvcencio/gotreesitter dependency-version: 0.47.0 dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <support@github.com>
dependabot
Bot
force-pushed
the
dependabot/go_modules/github.com/odvcencio/gotreesitter-0.47.0
branch
from
July 24, 2026 23:11
446476a to
a418068
Compare
dependabot
Bot
deleted the
dependabot/go_modules/github.com/odvcencio/gotreesitter-0.47.0
branch
July 24, 2026 23:15
townsendmerino
added a commit
that referenced
this pull request
Jul 25, 2026
…uspect Rework the campaign doc's problem statement to the three structural facts, adding the one that reframes the whole investigation: 3. The 256.6s bench ran ken v1.1.x (gotreesitter 0.20.5). The 0.20.x line silently disabled the hand-written repeat-boundary conflict resolvers for PHP (per the 0.21.0 changelog), causing GLR stack-forking blowups, and used pre-engine error recovery. We bumped to 0.47.0 on 2026-07-24 (PR #62) — the C-faithful engine + conflict-resolver fix + ~18% recovery throughput + non-quadratic cap-eviction. The bench predates the bump, so the parser is a prime suspect for ken being ~1.8x slower per file than the Python baseline. Adds an M0(a) sub-question + before/after table: index wall at v1.1.1 (0.20.5) vs HEAD (0.47.0), isolated to the dep, to quantify what the bump alone recovers. Notes the tension worth resolving: HEAD profiling STILL shows ~21% CPU in gotreesitter GLR retry on PHP, so the bump likely helped but didn't eliminate it. Also records the golden-drift risk (error-tree shapes changed in 0.21.0) that Task 2 audits, and reflects the M1 snapshot infra already landed (ADR-039). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
townsendmerino
added a commit
that referenced
this pull request
Jul 25, 2026
…e shape Task 2 of the cold-start campaign: audit PHP/tree-sitter goldens after the gotreesitter 0.20.5 → 0.47.0 bump (PR #62, 2026-07-24). Audit result: the full test suite is GREEN at 0.47.0 — no golden diffs to verify or regenerate. But every structural extractor test used deliberately well-formed source (the JS test comment even says so), so ken had ZERO malformed-input coverage for ANY language — a 0.21.0-style error-tree shape change (the changelog: PHP static named functions now explicitly named, conflict resolvers restored, C-faithful engine) could have ridden in unnoticed. This closes that gap for PHP, the language the 256.6 s bench exercised. Add TestBuild_PhpTruncatedController: a truncated Yii-style controller (cut off mid-expression, class body brace never closed) fed through Build. It pins the gotreesitter 0.47.0 error-recovery SHAPE as a bump guard: - robustness invariant (load-bearing): Build must not panic/error on truncated input, and the file is still indexed via error recovery; - complete method before the truncation (actionIndex) is recovered + named; - the truncated method (actionLogin) is NOT recovered as a function; - the unterminated class body yields no class; - clean imports before the truncation parse (Controller). Any future diff here means the grammar's error-tree shape changed — the test's doc comment directs the reader to verify against the changelog (C-oracle shapes) BEFORE re-baselining, per the campaign doc's golden-drift risk. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
townsendmerino
added a commit
that referenced
this pull request
Jul 25, 2026
…rse ~2.7x Task 3: timed `ken perf index --mode bm25` on a yii2 corpus (1,184 .php / 12,349 chunks), median of 5, temp binaries at each ref (GOWORK=off, own pins), M1 Pro / 16 GB, 2026-07-25. KEN_ENRICH=off is the control (removes the tree-sitter parse). enrich on : v1.1.1 (0.20.5) 1598ms vs HEAD (0.47.0) 3455ms = 2.16x slower enrich off: v1.1.1 537ms vs HEAD 557ms = ~equal The enrich-off control is flat across versions, so the ENTIRE on/off delta is the gotreesitter parse: the 0.47.0 C-faithful engine is ~2.7x slower on well-formed PHP than 0.20.5's conflict-resolver-disabled engine. This overturns fact (3)'s hypothesis. The 256.6s bench ran the FASTER parser; 0.20.5's GLR blowup is input-specific and yii2 doesn't trigger it enough to lose the well-formed-code speed — so the parser was not the bench bottleneck we suspected. And the correctness bump (PR #62) carries a real, previously unmeasured cold-start regression on typical PHP. Re-weights the campaign toward the caching levers (M1/M2 + the shipped lazy-structural win) and promotes "upstream the PHP parse cost" to a real target: get 0.47.0 correctness without its ~2.7x well-formed-PHP parse cost. M1 must land before the external rebench. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
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.
Bumps github.com/odvcencio/gotreesitter from 0.20.5 to 0.47.0.
Release notes
Sourced from github.com/odvcencio/gotreesitter's releases.
... (truncated)
Changelog
Sourced from github.com/odvcencio/gotreesitter's changelog.
... (truncated)
Commits
000ae6cMerge pull request #452 from odvcencio/codex/release-cadencef03fdbarelease(release): release v0.47.0 and document release cadencedc4cd83Merge pull request #451 from odvcencio/codex/stateful-checkpoint-contract15e1926fix(glr forest): restrict checkpoint capture to grammars with external scanners8c5d47cimprove(parser): Admit stateful scanners via exact checkpoint receiptscefdf91Merge pull request #449 from odvcencio/codex/gss-incremental-admission45b3d0aMerge pull request #450 from odvcencio/codex/issue429-scheduler-reuseb602584update(test): skip external scanner checkpoint fields in test167f40ftest: zero checkpoint slot counters in test comparison51ac420refactor(test): consolidate scanner unquiescent liveness verification