Replace compiler CPPO specializations with functors

[cristianoc]

Replace the compiler's six data-structure CPPO templates and their Dune generation rules with ordinary OCaml functors and small, explicit instantiation modules. Keep Hash_set_poly as a plain polymorphic implementation because its key type is not fixed by a functor argument.

Audit every resulting module against production consumers. Remove the unused Hash_set_int specialization and the effectively dead Hash_string specialization, whose only apparent compiler consumer was an always-empty cache for a commented-out function. Remove that cache and its uncalled Ext_ident.reset API; tests instantiate Hash.Make locally where a string-key table is useful.

Remove CPPO from the analysis and tools package dependencies. Retain it in the rescript package, with an explanatory comment, because the runtime's explicit regeneration script still invokes the external CPPO executable for its checked-in ReScript specializations.

This makes every compiler implementation visible to editors and build tooling, removes generated-source indirection, and uses the language's standard abstraction mechanism while preserving each retained specialization's comparison, equality, hashing, and printing semantics.

A four-way OCaml 5.3 release-build experiment compared concrete CPPO and functor implementations, both with and without Flambda. No meaningful compiler-speed difference was observed between CPPO and functors on clean builds of the 150-module runtime corpus, parallel builds, or incremental rebuilds; generated JavaScript was identical. Isolated data-structure benchmarks did show differences without Flambda, but those differences disappeared with Flambda.

Flambda itself was slower for representative compilation workloads while improving several sustained parsing and printing benchmarks. Its larger binary and startup cost appear to contribute to the typical-build slowdown, although startup-only measurements do not explain all of it. Investigating Flambda further remains worthwhile, particularly together with a persistent compiler server that can amortize per-file process startup.

Add direct coverage for the ordered local-identifier map and retain the existing map, set, hash, hash-set, and vector tests.

Signed-off-by: Cristiano Calcagno cristianoc@users.noreply.github.com

---

_{Stack created with GitHub Stacks CLI • Give Feedback 💬}

[codecov]

Codecov Report

❌ Patch coverage is 77.67176% with 117 lines in your changes missing coverage. Please review.
✅ Project coverage is 74.83%. Comparing base (9f26fa6) to head (c9fa8c5).
⚠️ Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
compiler/ext/vec.ml	67.94%	50 Missing ⚠️
compiler/ext/ext_map.ml	66.66%	31 Missing ⚠️
compiler/ext/hash.ml	68.33%	19 Missing ⚠️
compiler/ext/ordered_hash_map.ml	76.74%	10 Missing ⚠️
compiler/ext/hash_set.ml	77.27%	5 Missing ⚠️
compiler/ext/ext_set.ml	97.95%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #8542      +/-   ##
==========================================
+ Coverage   74.70%   74.83%   +0.13%     
==========================================
  Files         460      477      +17     
  Lines       62389    63311     +922     
==========================================
+ Hits        46608    47380     +772     
- Misses      15781    15931     +150     

Files with missing lines	Coverage Δ
compiler/ext/ext_ident.ml	72.72% <ø> (+0.66%)	⬆️
compiler/ext/hash_ident.ml	100.00% <100.00%> (ø)
compiler/ext/hash_set_ident.ml	100.00% <100.00%> (ø)
compiler/ext/hash_set_poly.ml	100.00% <100.00%> (ø)
compiler/ext/ordered_hash_map_local_ident.ml	100.00% <100.00%> (ø)
compiler/ext/set_ident.ml	100.00% <100.00%> (ø)
tests/ounit_tests/ounit_hash_stubs_test.ml	100.00% <ø> (+28.57%)	⬆️
tests/ounit_tests/ounit_hashtbl_tests.ml	100.00% <100.00%> (ø)
tests/ounit_tests/ounit_scc_tests.ml	84.33% <100.00%> (ø)
compiler/ext/ext_set.ml	97.95% <97.95%> (ø)
... and 5 more

... and 4 files with indirect coverage changes

🚀 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@8542

@rescript/darwin-arm64

npm i https://pkg.pr.new/@rescript/darwin-arm64@8542

@rescript/darwin-x64

npm i https://pkg.pr.new/@rescript/darwin-x64@8542

@rescript/linux-arm64

npm i https://pkg.pr.new/@rescript/linux-arm64@8542

@rescript/linux-x64

npm i https://pkg.pr.new/@rescript/linux-x64@8542

@rescript/runtime

npm i https://pkg.pr.new/@rescript/runtime@8542

@rescript/win32-x64

npm i https://pkg.pr.new/@rescript/win32-x64@8542

commit: c9fa8c5

[cknitt]

@codex review

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. More of your lovely PRs please.

Reviewed commit: 6beded4738

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[github-actions]

Developer playground preview: https://rescript-lang.github.io/rescript/dev-playground/?version=pr-8542