[cknitt]

Part of #9256 (duplicate check of pattern variables).

Typecore.enter_variable compared each new variable with all variables of the pattern so far, and a let rec … and … group is one pattern. The bound names are now kept in a set next to the list. Both live in one value of the request state, so resetting the pattern, add_pattern_variables and the save and restore around or-patterns carry the set along without extra code.

Measured with bsc on linux-arm64, a let rec group of n bindings (allocations were already linear; this saves string comparisons):

Bindings	Before	After
4,000	514 ms	475 ms
8,000	1.14 s	0.97 s
16,000	2.69 s	1.96 s

New tests: multiply_bound_variable_let_rec (a duplicate in a let rec group), multiply_bound_variable_or_pattern (an or-pattern binding the same name on both sides is accepted, then a real duplicate is reported) and multiply_bound_variable_alias (Some(a) as a).

🤖 Generated with Claude Code

https://claude.ai/code/session_01Kbf3gK1dcjbe97yY1wRMgs

[github-actions]

Changelog

💅 Polish

• Speed up type checking of patterns that bind many variables, such as large let rec groups.

Semantic Diff

1 commit · Full page

Find duplicate pattern variables in a set of the bound names

_{4bae89d · cknitt + Claude · Nov 07, 2026, 10:07 UTC}

Why

Each new pattern variable was compared with every variable the pattern had bound so far, and a let rec … and … group is one pattern, so n bindings cost O(n²) string comparisons: 16,000 bindings took 2.7 s. They now take 2.0 s.

The names are kept in a set next to the list of variables, in one value, so that resetting the pattern and the save and restore around or-patterns carry both.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 81.74%. Comparing base (bf35764) to head (4bae89d).

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #9260      +/-   ##
==========================================
- Coverage   81.74%   81.74%   -0.01%     
==========================================
  Files         457      457              
  Lines       65713    65712       -1     
==========================================
- Hits        53715    53714       -1     
  Misses      11998    11998              

Files with missing lines	Coverage Δ
compiler/ml/typecore.ml	89.11% <100.00%> (-0.01%)	⬇️

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[pkg-pr-new]

Open in StackBlitz

rescript

npm i https://pkg.pr.new/rescript@9260

@rescript/belt

npm i https://pkg.pr.new/@rescript/belt@9260

@rescript/darwin-arm64

npm i https://pkg.pr.new/@rescript/darwin-arm64@9260

@rescript/linux-arm64

npm i https://pkg.pr.new/@rescript/linux-arm64@9260

@rescript/linux-x64

npm i https://pkg.pr.new/@rescript/linux-x64@9260

@rescript/runtime

npm i https://pkg.pr.new/@rescript/runtime@9260

@rescript/win32-x64

npm i https://pkg.pr.new/@rescript/win32-x64@9260

commit: 4bae89d

[github-actions]

Developer playground preview: https://rescript-lang.github.io/rescript/dev-playground/?version=pr-9260

[github-actions]

🎉 Compiler benchmarks

Allocated words

In total: 456,989,425 → 457,158,125 (+0.04%).

Benchmark	Base	Head	Change
stress/LetRec.res	35,120,632	35,257,343	+0.39%

Instructions

In total: 10,226,962,259 → 10,126,222,706 (-0.99%).

🎉 Improvement: 1 benchmark executes more than 1% fewer instructions.

Benchmark	Base	Head	Change
stress/LetRec.res	1,431,510,943	1,325,636,900	-7.40%
stress/ObjectFields.res	957,097,722	962,171,555	+0.53%

Pull request 4bae89d compared with base bf35764 in this run.