Make Lambda-to-Lam conversion structural - #8604
Merged
Merged
Conversation
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 9762eed607
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".
Pattern matching introduces Alias lets as O(1) renames of a pattern ident onto the scrutinee. Convert used to substitute those while translating; do the same rewrite as a Lam pass immediately after conversion instead, so convert stays a translation. Exported aliases are kept for coercion. Later lets_dce is unchanged. JS from runtime, Belt, tests/tests, belt_tests, and commonjs_tests matched the convert-time pass (1079 files). Signed-off-by: Cristiano Calcagno <cristianoc@users.noreply.github.com>
The Lam primitive was compiled to f.apply(null, args) but never constructed. Drop it together with the JS IR constructor and the stale Lifused/translclass comments left in convert. Signed-off-by: Cristiano Calcagno <cristianoc@users.noreply.github.com>
%revapply/%apply were leftover |>/@@ encodings; ReScript -> is rewritten before typing. __LOC__ and friends still compile, but as constants in translcore rather than a Lambda primitive. Signed-off-by: Cristiano Calcagno <cristianoc@users.noreply.github.com>
Pack JS catch values in translcore so convert only translates Ltrywith. Drop raise_kind/reraise tracking (JS throw is one operation) and emit RE_EXN_ID string equality from matching instead of Pextension_slot_eq. Signed-off-by: Cristiano Calcagno <cristianoc@users.noreply.github.com>
Exception packing, Promise.catch, and JsExn.anyToExnInternal now apply Primitive_exceptions.internalToException instead of a dedicated primitive. Signed-off-by: Cristiano Calcagno <cristianoc@users.noreply.github.com>
Emit nested let/letrec from transl_let and eval_rec_bindings via Lambda_scc.bind_rec. Convert Lletrec is identity. JS compile only sorts functions before values for dummy/updateDummy init. Signed-off-by: Cristiano Calcagno <cristianoc@users.noreply.github.com>
Remove the dense int-switch-to-add peephole and the switcher-offset let rewrite. Rewrite the one test that relied on the former to an or-pattern plus add. Signed-off-by: Cristiano Calcagno <cristianoc@users.noreply.github.com>
Drop exit_map and the catch-of-raise peephole. Matching still emits catch body with (i) (exit j); Lam_pass_exits inlines the size-1 handler. Signed-off-by: Cristiano Calcagno <cristianoc@users.noreply.github.com>
Signed-off-by: Cristiano Calcagno <cristianoc@users.noreply.github.com>
Lambda.mk_prim expands Peliminated so those payloads never appear as Lprim nodes. Convert maps the constructor 1-1. Lam matches assert false because the node cannot occur. Signed-off-by: Cristiano Calcagno <cristianoc@users.noreply.github.com>
Replace Lambda Const_base and Const_immstring with direct int, char, string, float, and bigint constructors, and use int32-backed Const_int consistently in Lambda and Lam. Represent Lam assert-false separately from ordinary integer zero. Remove the unreachable typedtree Const_int32 and Const_int64 variants and bump the CMT magic number for the serialized type change. Treat ReScript strings as immutable when normalizing match actions, with snapshots covering early string-action sharing and the separate untagged-discriminator merge boundary. Signed-off-by: Cristiano Calcagno <cristianoc@users.noreply.github.com>
Give Lambda Const_string the same structured payload used by Lam, and move processed-delimiter decoding into External_arg_spec so Lambda producers can store the final delimiter directly. Remove the inline-constant delimiter round-trip, cover every processed delimiter encoding, and consolidate the branch changelog into review-facing entries. Signed-off-by: Cristiano Calcagno <cristianoc@users.noreply.github.com>
Signed-off-by: Cristiano Calcagno <cristianoc@users.noreply.github.com>
cristianoc
force-pushed
the
lam-convert-cleanup
branch
from
August 31, 2026 13:50
9762eed to
262cd0b
Compare
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #8604 +/- ##
==========================================
+ Coverage 76.33% 76.44% +0.11%
==========================================
Files 476 478 +2
Lines 63253 63163 -90
==========================================
+ Hits 48284 48286 +2
+ Misses 14969 14877 -92
🚀 New features to boost your workflow:
|
rescript
@rescript/belt
@rescript/darwin-arm64
@rescript/darwin-x64
@rescript/linux-arm64
@rescript/linux-x64
@rescript/runtime
@rescript/win32-x64
commit: |
cknitt
reviewed
Aug 31, 2026
| *) | ||
|
|
||
| (* Semantic SCC already ran in [Lambda_scc.bind_rec]. JS still wants | ||
| functions before values so dummy / updateDummy init is well-ordered. *) |
Member
There was a problem hiding this comment.
what is dummy / updateDummy init?
Collaborator
Author
There was a problem hiding this comment.
it refers to the code below -- can clean up the comment
|
Developer playground preview: https://rescript-lang.github.io/rescript/dev-playground/?version=pr-8604 |
cknitt
approved these changes
Aug 31, 2026
cknitt
left a comment
cknitt
left a comment
Member
There was a problem hiding this comment.
In a large project, I got three output changes. Two were improvements (unnecessary local vars removed/inlined).
One regarding two branches in a switch statement with identical body was a bit unexpected (introducing an exit variable that was not needed before), but I think that particular case can be investigated separately later.
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
2 participants
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.
Summary
Lam_sccThis PR is based on #8598 so the refreshed compiler documentation describes the contracts used by this cleanup. It is intentionally not registered as a GitHub stack.
Generated output audit
Every generated JavaScript change was reviewed:
internalToExceptionimplementation through the normal runtime binding after removal ofPwrap_exnConst_stringdelimiter change itself produces no generated JavaScript changesTests
make test