Bump github.com/odvcencio/gotreesitter from 0.53.0 to 0.55.1

[dependabot]

Bumps github.com/odvcencio/gotreesitter from 0.53.0 to 0.55.1.

Release notes

Sourced from github.com/odvcencio/gotreesitter's releases.

gotreesitter v0.55.1

Fixed

• Restore the Python initial stack cap of two when the source has no possible unpacking expression. Keep eight stacks for possible unpacking and preserve explicit stack overrides (#1290). The report's inline generator builds 371,146 work nodes instead of 623,912. Production fresh parsing fell from 267.331 to 161.811 ms; the tree retains 59,105 nodes. The seven list-splat fixtures still match locked C. Django retains six differences across 2,932 files.
• Preserve conflict forks beyond stack depth 4,096 and reset the dispatch guard when reductions reach a new minimum depth (#1291). C now returns clean, complete trees for 4,091 functions and up to 65,536 string calls in one function. Both lexers pass fresh, incremental, and locked C comparisons at the tested counts. At 8,192 calls, TokenSource parsing fell from 2,445.8 to 115.7 ms. The deterministic finite automaton (DFA) lexer fell from 2,425.1 to 123.7 ms.
• Remove a duplicate clone during hidden-node alias materialization (#1292). Scala work nodes fell from 168,685 to 149,592 on both routes; the tree retains 57,281 nodes. Production fresh parsing fell from 73.565 to 70.638 ms. Work remains 4.5% above the old grammar's 143,197 nodes.
• Extend the poor-yield reuse guard to roots with more than four children (#1292). A Dart proxy fixture now retries fresh for expensive edits and retains profitable suffix reuse. Production insertion fell from 240.990 to 123.209 ms; deletion fell from 718.243 to 123.759 ms.

Documentation

• Align the roadmap, repository map, and agent workflow with the owner's v1 design (#1329). This patch adds no public API or admission-default change.

Measurement scope

The linked pull requests contain separate correctness and performance evidence. Each completed the randomized Go benchmark trio with 20 shuffle seeds and 750 ms per case. The runs used one process per seed and GOMAXPROCS=1. No primary timing change was significant; primary allocations stayed at eight, five, and zero per operation. These Linux measurements do not establish Windows performance. Python uses the reporter's inline generator. Dart uses a 137 KiB proxy with 45.4% baseline reuse. The reporter measured 41.8% Dart reuse; the exact fixture remains unverified. Scala compact timing varied, with some medians increasing despite fewer work nodes. The standard Scala grammar importer still rejects _end_marker_named_tail; focused locked C checks pass. C corpus results remain 23/25 clean files and 20/25 deep matches.

Publication exception

On 2026-09-26, the owner authorized the v0.55.1-only tag-creation exception. The exception permits publication without a workflow-only tag-creation actor rule. All other gates remain mandatory under the release process. This urgent patch addresses the Python regression and false C errors reported in #454.

... (truncated)

Changelog

Sourced from github.com/odvcencio/gotreesitter's changelog.

[0.55.1] - 2026-09-26

Fixed

• Restore the Python initial stack cap of two when the source has no possible unpacking expression. Keep eight stacks for possible unpacking and preserve explicit stack overrides (#1290). The report's inline generator builds 371,146 work nodes instead of 623,912. Production fresh parsing fell from 267.331 to 161.811 ms; the tree retains 59,105 nodes. The seven list-splat fixtures still match locked C. Django retains six differences across 2,932 files.
• Preserve conflict forks beyond stack depth 4,096 and reset the dispatch guard when reductions reach a new minimum depth (#1291). C now returns clean, complete trees for 4,091 functions and up to 65,536 string calls in one function. Both lexers pass fresh, incremental, and locked C comparisons at the tested counts. At 8,192 calls, TokenSource parsing fell from 2,445.8 to 115.7 ms. The deterministic finite automaton (DFA) lexer fell from 2,425.1 to 123.7 ms.
• Remove a duplicate clone during hidden-node alias materialization (#1292). Scala work nodes fell from 168,685 to 149,592 on both routes; the tree retains 57,281 nodes. Production fresh parsing fell from 73.565 to 70.638 ms. Work remains 4.5% above the old grammar's 143,197 nodes.
• Extend the poor-yield reuse guard to roots with more than four children (#1292). A Dart proxy fixture now retries fresh for expensive edits and retains profitable suffix reuse. Production insertion fell from 240.990 to 123.209 ms; deletion fell from 718.243 to 123.759 ms.

Documentation

• Align the roadmap, repository map, and agent workflow with the owner's v1 design (#1329). This patch adds no public API or admission-default change.

Measurement scope

The linked pull requests contain separate correctness and performance evidence. Each completed the randomized Go benchmark trio with 20 shuffle seeds and 750 ms per case. The runs used one process per seed and GOMAXPROCS=1. No primary timing change was significant; primary allocations stayed at eight, five, and zero per operation. These Linux measurements do not establish Windows performance. Python uses the reporter's inline generator. Dart uses a 137 KiB proxy with 45.4% baseline reuse. The reporter measured 41.8% Dart reuse; the exact fixture remains unverified. Scala compact timing varied, with some medians increasing despite fewer work nodes. The standard Scala grammar importer still rejects _end_marker_named_tail; focused locked C checks pass. C corpus results remain 23/25 clean files and 20/25 deep matches.

Publication exception

On 2026-09-26, the owner authorized the v0.55.1-only tag-creation exception. The exception permits publication without a workflow-only tag-creation actor rule. All other gates remain mandatory under the release process.

... (truncated)

Commits

• 92db945 update(release): document the v0.55.1 release and its exception (#1330)
• ad0fed1 update(docs): align repository guidance with the v1 design (#1329)
• 10834e2 fix(parser): reduce alias allocations and bound Dart reuse (#1292)
• c3ae5dc Restore Python parse speed on files without splats (#1290)
• e436c82 fix(parser): keep GLR conflict forks past deep stacks (#1291)
• 849e775 document(docs): document v0.55.0 release evidence and tag exception (#1289)
• 914cb71 fix(recovery): bound EOF recovery versions to stop runaway memory (#1287)
• d25623f add(parsercore): elect recovered material roots in the winning recovery group...
• 7278877 add(compact): record peak headers and derivations for certification (#1285)
• 170ee57 fix(parsercore): stop treating accept as shiftable in S5 collector (#1284)
• Additional commits viewable in compare view

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

---

Dependabot commands and options

You can trigger Dependabot actions by commenting on this PR:

• @dependabot rebase will rebase this PR
• @dependabot recreate will recreate this PR, overwriting any edits that have been made to it
• @dependabot show <dependency name> ignore conditions will show all of the ignore conditions of the specified dependency
• @dependabot ignore this major version will close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)
• @dependabot ignore this minor version will close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)
• @dependabot ignore this dependency will close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself)