BREAKING CHANGE: Rescript 12 support #96

illusionalsagacity · 2025-12-02T01:43:02Z

The output type has changed from Js.Json.t to JSON.t
Adds handling for result in addition to the existing Belt.Result.t type
Polyvariant tags no longer require the intermediate Js.Json.tagged type
There is still some usage of Belt.Array.getExn in the PPX generated code, is this ok to change to Array.getUnsafe?
Set up pnpm workspaces so examples and test are automatically linked to the parent folder
needs testing on other codebases

mununki · 2025-12-02T01:57:27Z

Thank you for the PR!

There is still some usage of Belt.Array.getExn in the PPX generated code, is this ok to change to Array.getUnsafe?

Okay!

illusionalsagacity · 2025-12-02T02:49:19Z

Thank you for the PR!

There is still some usage of Belt.Array.getExn in the PPX generated code, is this ok to change to Array.getUnsafe?

Okay!

oh by the way, do you have an easy way to publish pre-releases on npm? Preferably I'd like to test this against a larger codebase I'm using spice for before merging

- [x] Js.Json.t type has been replaced with JSON.t - [x] Polyvariant tags no longer require the intermediate Js.Json.tagged type, removing some extra function calls - [x] Usage of Js and Belt in the PPX generated code has been migrated to Stdlib - [x] Updated examples and tests to reflect the above changes

mununki · 2025-12-02T04:04:41Z

oh by the way, do you have an easy way to publish pre-releases on npm? Preferably I'd like to test this against a larger codebase I'm using spice for before merging

Sounds good to me. Let me figure it out.

package.json

CHANGELOG.md

mununki · 2025-12-03T01:28:48Z

@illusionalsagacity The 0.3.0-next.0 is published to the npm now.

illusionalsagacity · 2025-12-03T01:53:46Z

@illusionalsagacity The 0.3.0-next.0 is published to the npm now.

Thanks! unfortunately I ran into issue #85 when trying to test it

Depends on green-labs#96 - Introduced a comprehensive benchmark file (Benchmark.res) containing various types including records, variants, polyvariants, and tuples to evaluate ppx performance. - Added types with optional fields, nested records, and large variants to stress test the system. - Created utility functions in NoSpiceAnnotations.mjs for user management and status checking without spice annotations.

illusionalsagacity · 2025-12-03T04:06:29Z

I'm also getting a very strange stalling/hanging with rewatch after it gets pretty far into the compilation, and with rescript-legacy I get the following:

rescript: [1/6] src/rescript/Spice.ast
FAILED: src/rescript/Spice.ast
Ill-formed list of warnings
rescript: [2/6] src/rescript/Spice_Codecs.ast
FAILED: src/rescript/Spice_Codecs.ast
Ill-formed list of warnings
FAILED: cannot make progress due to previous errors.

I'm not sure if the compiler performance is related to this, or just the "antivirus" on my work laptop going crazy.

mununki · 2025-12-03T06:01:29Z

I'm also getting a very strange stalling/hanging with rewatch after it gets pretty far into the compilation, and with rescript-legacy I get the following:
rescript: [1/6] src/rescript/Spice.ast
FAILED: src/rescript/Spice.ast
Ill-formed list of warnings
rescript: [2/6] src/rescript/Spice_Codecs.ast
FAILED: src/rescript/Spice_Codecs.ast
Ill-formed list of warnings
FAILED: cannot make progress due to previous errors.
I'm not sure if the compiler performance is related to this, or just the "antivirus" on my work laptop going crazy.

Hmm.. Looks quite odd to me. Let me investigate little further about the error message Ill-formed list of warnings regarding the recent changes of the compiler.

Depends on green-labs#96 - Introduced a comprehensive benchmark file (Benchmark.res) containing various types including records, variants, polyvariants, and tuples to evaluate ppx performance. - Added types with optional fields, nested records, and large variants to stress test the system. - Created utility functions in NoSpiceAnnotations.mjs for user management and status checking without spice annotations.

illusionalsagacity · 2025-12-06T22:47:16Z

@mununki I was able to get this compiling on my codebase, but I needed to incorporate #98 because with just this change alone, the compiler would hang indefinitely with some of the record types we had with ~50 fields. I also found #99 in some of our Api tests.

mununki · 2025-12-09T05:45:39Z

@mununki I was able to get this compiling on my codebase, but I needed to incorporate #98 because with just this change alone, the compiler would hang indefinitely with some of the record types we had with ~50 fields. I also found #99 in some of our Api tests.

#98 is a known issue. Can we separate the #98 work from merging this PR first?

mununki · 2025-12-09T05:54:21Z

test/test/__tests__/spec/variants_test.mjs

 Zora.test("variants with @spice.as", t => {
  let variantEncoded = Variants.t_encode("One");
-  testEqual(t, "encode 하나", variantEncoded, 1);
+  testEqual(t, `encode 하나`, variantEncoded, 1);


nit: It's not a big deal, but wondering why " is changed to `?

It looks like on v12 the compiler now emits backtick strings if that how it was written in the .res file

https://rescript-lang.org/try/?version=v11.1.4&module=esmodule&code=DYUwLgBAZg9jEF4IANY2QKFJARgQwCdEIAifAkrcaOc4kgHTLwC8ByEoA

Oh, I get it now. Thank you for explaining it.

mununki · 2025-12-09T05:54:48Z

examples/rescript.json

@@ -1,5 +1,5 @@
 {
-  "name": "ppx-spice-examples",
+  "name": "spice-ppx-examples",


nit: need to change the name?

This was giving me a warning from the new rewatch build system, something about breaking package resolution

I get it now. Nevermind, I'll fix these names later.

illusionalsagacity · 2025-12-09T06:01:54Z

@mununki I was able to get this compiling on my codebase, but I needed to incorporate #98 because with just this change alone, the compiler would hang indefinitely with some of the record types we had with ~50 fields. I also found #99 in some of our Api tests.

#98 is a known issue. Can we separate the #98 work from merging this PR first?

Ah, by 'incorporate' I meant on the build I was using to test; but yeah I can untangle the changes in #98 from the changes here

mununki · 2025-12-09T06:03:20Z

@mununki I was able to get this compiling on my codebase, but I needed to incorporate #98 because with just this change alone, the compiler would hang indefinitely with some of the record types we had with ~50 fields. I also found #99 in some of our Api tests.

#98 is a known issue. Can we separate the #98 work from merging this PR first?

Ah, by 'incorporate' I meant on the build I was using to test; but yeah I can untangle the changes in #98 from the changes here

Thanks!

mununki · 2025-12-09T06:04:51Z

I think that @rescript/runtime package is needed to run the test in /test. Don't we need to add it as a dev dependency?

test git:(illusionalsagacity-rescript-12-support) ✗ pnpm test      

> spice-ppx-test@1.0.0 test /Users/mununki/GitHub/gl/ppx_spice/test
> pta './test/__tests__/Index.mjs' | tap-difflet

node:internal/modules/package_json_reader:316
  throw new ERR_MODULE_NOT_FOUND(packageName, fileURLToPath(base), null);
        ^

Error [ERR_MODULE_NOT_FOUND]: Cannot find package '@rescript/runtime' imported from /Users/mununki/GitHub/gl/ppx_spice/test/src/Records.mjs
    at Object.getPackageJSONURL (node:internal/modules/package_json_reader:316:9)
    at packageResolve (node:internal/modules/esm/resolve:768:81)
    at moduleResolve (node:internal/modules/esm/resolve:858:18)
    at defaultResolve (node:internal/modules/esm/resolve:990:11)
    at #cachedDefaultResolve (node:internal/modules/esm/loader:737:20)
    at ModuleLoader.resolve (node:internal/modules/esm/loader:714:38)
    at ModuleLoader.getModuleJobForImport (node:internal/modules/esm/loader:293:38)
    at #link (node:internal/modules/esm/module_job:208:49)
    at process.processTicksAndRejections (node:internal/process/task_queues:103:5) {
  code: 'ERR_MODULE_NOT_FOUND'
}

Node.js v24.11.1

illusionalsagacity

I think that @rescript/runtime package is needed to run the test in /test. Don't we need to add it as a dev dependency?

test git:(illusionalsagacity-rescript-12-support) ✗ pnpm test      

> spice-ppx-test@1.0.0 test /Users/mununki/GitHub/gl/ppx_spice/test
> pta './test/__tests__/Index.mjs' | tap-difflet

node:internal/modules/package_json_reader:316
  throw new ERR_MODULE_NOT_FOUND(packageName, fileURLToPath(base), null);
        ^

Error [ERR_MODULE_NOT_FOUND]: Cannot find package '@rescript/runtime' imported from /Users/mununki/GitHub/gl/ppx_spice/test/src/Records.mjs
    at Object.getPackageJSONURL (node:internal/modules/package_json_reader:316:9)
    at packageResolve (node:internal/modules/esm/resolve:768:81)
    at moduleResolve (node:internal/modules/esm/resolve:858:18)
    at defaultResolve (node:internal/modules/esm/resolve:990:11)
    at #cachedDefaultResolve (node:internal/modules/esm/loader:737:20)
    at ModuleLoader.resolve (node:internal/modules/esm/loader:714:38)
    at ModuleLoader.getModuleJobForImport (node:internal/modules/esm/loader:293:38)
    at #link (node:internal/modules/esm/module_job:208:49)
    at process.processTicksAndRejections (node:internal/process/task_queues:103:5) {
  code: 'ERR_MODULE_NOT_FOUND'
}

Node.js v24.11.1

Hmm, that is odd. It's listed in the pnpm lockfile and yarn sees it:

$ pnpm why @rescript/runtime
Legend: production dependency, optional only, dev only

spice-ppx-test@1.0.0 /Users/rob/src/github.com/illusionalsagacity/ppx_spice/test

devDependencies:
rescript 12.0.0
└── @rescript/runtime 12.0.0

$ yarn why @rescript/runtime
yarn why v1.22.22
[1/4] 🤔  Why do we have the module "@rescript/runtime"...?
[2/4] 🚚  Initialising dependency graph...
[3/4] 🔍  Finding dependency...
[4/4] 🚡  Calculating file sizes...
=> Found "@rescript/runtime@12.0.0"
info Reasons this module exists
   - "rescript" depends on it
   - Hoisted from "rescript#@rescript#runtime"
info Disk size without dependencies: "19.2MB"
info Disk size with unique dependencies: "19.2MB"
info Disk size with transitive dependencies: "19.2MB"
info Number of shared dependencies: 0

mununki · 2025-12-09T06:20:57Z

Hmm, that is odd. It's listed in the pnpm lockfile and yarn sees it:

$ pnpm why @rescript/runtime
Legend: production dependency, optional only, dev only

spice-ppx-test@1.0.0 /Users/rob/src/github.com/illusionalsagacity/ppx_spice/test

devDependencies:
rescript 12.0.0
└── @rescript/runtime 12.0.0

$ yarn why @rescript/runtime
yarn why v1.22.22
[1/4] 🤔  Why do we have the module "@rescript/runtime"...?
[2/4] 🚚  Initialising dependency graph...
[3/4] 🔍  Finding dependency...
[4/4] 🚡  Calculating file sizes...
=> Found "@rescript/runtime@12.0.0"
info Reasons this module exists
   - "rescript" depends on it
   - Hoisted from "rescript#@rescript#runtime"
info Disk size without dependencies: "19.2MB"
info Disk size with unique dependencies: "19.2MB"
info Disk size with transitive dependencies: "19.2MB"
info Number of shared dependencies: 0

What is your $PWD running this commands?

mununki · 2025-12-09T06:24:07Z

Hmm, that is odd. It's listed in the pnpm lockfile and yarn sees it:

$ pnpm why @rescript/runtime
Legend: production dependency, optional only, dev only

spice-ppx-test@1.0.0 /Users/rob/src/github.com/illusionalsagacity/ppx_spice/test

devDependencies:
rescript 12.0.0
└── @rescript/runtime 12.0.0

$ yarn why @rescript/runtime
yarn why v1.22.22
[1/4] 🤔  Why do we have the module "@rescript/runtime"...?
[2/4] 🚚  Initialising dependency graph...
[3/4] 🔍  Finding dependency...
[4/4] 🚡  Calculating file sizes...
=> Found "@rescript/runtime@12.0.0"
info Reasons this module exists
   - "rescript" depends on it
   - Hoisted from "rescript#@rescript#runtime"
info Disk size without dependencies: "19.2MB"
info Disk size with unique dependencies: "19.2MB"
info Disk size with transitive dependencies: "19.2MB"
info Number of shared dependencies: 0

What is your $PWD running this commands?

It works fine after cleaning the node_modules dir.

mununki

Thank you for your work!

mununki · 2025-12-09T06:25:09Z

Any thing left behind this PR?

Depends on green-labs#96 - Introduced a comprehensive benchmark file (Benchmark.res) containing various types including records, variants, polyvariants, and tuples to evaluate ppx performance. - Added types with optional fields, nested records, and large variants to stress test the system. - Created utility functions in NoSpiceAnnotations.mjs for user management and status checking without spice annotations.

illusionalsagacity · 2025-12-09T15:13:55Z

Any thing left behind this PR?

I don't think so, did you want to merge this one first and then #98?

mununki · 2025-12-10T00:47:58Z

Any thing left behind this PR?

I don't think so, did you want to merge this one first and then #98?

I'd like to merge this PR first, then performance improvement such as #98 if you don't mind. Therefore, can you separate this commit illusionalsagacity@5177e87 to another PR? I'd like to minimize this PR and focus to the support the compiler v12.

illusionalsagacity · 2025-12-10T04:34:12Z

Any thing left behind this PR?

I don't think so, did you want to merge this one first and then #98?

I'd like to merge this PR first, then performance improvement such as #98 if you don't mind. Therefore, can you separate this commit illusionalsagacity@5177e87 to another PR? I'd like to minimize this PR and focus to the support the compiler v12.

I'm not seeing that commit in this PR, it's in the perf-improvement branch

mununki · 2025-12-10T06:15:44Z

Oh, I'm confused about the GitHub ui.

benadamstyles · 2025-12-12T15:23:13Z

@illusionalsagacity I'm also getting Ill-formed list of warnings when using rescript-legacy. Did you find a solution?

illusionalsagacity · 2025-12-12T15:28:48Z

@illusionalsagacity I'm also getting Ill-formed list of warnings when using rescript-legacy. Did you find a solution?

I did not, the new build system ended up working for me with 12.0.1 and the changes in #98

benadamstyles · 2025-12-12T15:44:38Z

@illusionalsagacity Thanks for the reply. Unfortunately we can't use the new build system.

I think that Spice cannot be used at all in ReScript v12 with the legacy build system. Am I correct, or do you know of some version that does work in these circumstances?

mununki · 2025-12-13T07:25:16Z

@illusionalsagacity Thanks for the reply. Unfortunately we can't use the new build system.

I think that Spice cannot be used at all in ReScript v12 with the legacy build system. Am I correct, or do you know of some version that does work in these circumstances?

Can you provide a reproducible project for debugging?

benadamstyles · 2025-12-15T12:12:26Z

@mununki Of course, here you go: https://github.com/benadamstyles/rescript-v12-legacy-build-spice-repro

benadamstyles · 2025-12-16T15:02:38Z

I've added an issue #109

mununki · 2026-01-09T01:30:10Z

test/rescript.json

-    "../ppx"
-  ],
+  "dev-dependencies": ["@dusty-phillips/rescript-zora"],
+  "ppx-flags": ["@greenlabs/ppx-spice/ppx"],


I seem to have made a mistake in my review. It seems like ppx-flags should be handled by the locally built ppx executable binary. I'll correct this in a separate PR.

I seem to have made a mistake in my review. It seems like ppx-flags should be handled by the locally built ppx executable binary. I'll correct this in a separate PR.

My bad. I found out that the ppx is added as a workspace dependency.

illusionalsagacity force-pushed the rescript-12-support branch from d890744 to 37c1608 Compare December 2, 2025 03:15

mununki reviewed Dec 2, 2025

View reviewed changes

package.json Outdated Show resolved Hide resolved

illusionalsagacity mentioned this pull request Dec 2, 2025

Optional type encodes to option<Js.Json.t> instead of Js.Json.t #85

Open

mununki reviewed Dec 3, 2025

View reviewed changes

CHANGELOG.md Outdated Show resolved Hide resolved

illusionalsagacity force-pushed the rescript-12-support branch from c77eafd to c55e055 Compare December 3, 2025 01:48

chore: update version and changelog

8d77389

illusionalsagacity force-pushed the rescript-12-support branch from c55e055 to 8d77389 Compare December 3, 2025 01:49

illusionalsagacity mentioned this pull request Dec 3, 2025

Improve algorithmic complexity of AST generation #98

Open

mununki reviewed Dec 9, 2025

View reviewed changes

illusionalsagacity commented Dec 9, 2025

View reviewed changes

mununki approved these changes Dec 9, 2025

View reviewed changes

mununki merged commit 75ba4af into green-labs:main Dec 10, 2025
2 checks passed

illusionalsagacity deleted the rescript-12-support branch December 10, 2025 07:40

illusionalsagacity mentioned this pull request Dec 13, 2025

PPX errors with "Ill-formed list of warnings" after removal of -bs-v flag rescript-lang/rescript#8090

Closed

5 tasks

mununki reviewed Jan 9, 2026

View reviewed changes

BREAKING CHANGE: Rescript 12 support #96

BREAKING CHANGE: Rescript 12 support #96

Uh oh!

Conversation

illusionalsagacity commented Dec 2, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

mununki commented Dec 2, 2025

Uh oh!

illusionalsagacity commented Dec 2, 2025

Uh oh!

mununki commented Dec 2, 2025

Uh oh!

Uh oh!

Uh oh!

mununki commented Dec 3, 2025

Uh oh!

illusionalsagacity commented Dec 3, 2025

Uh oh!

illusionalsagacity commented Dec 3, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

mununki commented Dec 3, 2025

Uh oh!

illusionalsagacity commented Dec 6, 2025

Uh oh!

mununki commented Dec 9, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

illusionalsagacity commented Dec 9, 2025

Uh oh!

mununki commented Dec 9, 2025

Uh oh!

mununki commented Dec 9, 2025

Uh oh!

illusionalsagacity left a comment

Choose a reason for hiding this comment

Uh oh!

mununki commented Dec 9, 2025

Uh oh!

mununki commented Dec 9, 2025

Uh oh!

mununki left a comment

Choose a reason for hiding this comment

Uh oh!

mununki commented Dec 9, 2025

Uh oh!

illusionalsagacity commented Dec 9, 2025

Uh oh!

mununki commented Dec 10, 2025

Uh oh!

illusionalsagacity commented Dec 10, 2025

Uh oh!

mununki commented Dec 10, 2025

Uh oh!

Uh oh!

benadamstyles commented Dec 12, 2025

Uh oh!

illusionalsagacity commented Dec 12, 2025

Uh oh!

benadamstyles commented Dec 12, 2025

Uh oh!

mununki commented Dec 13, 2025

Uh oh!

benadamstyles commented Dec 15, 2025

Uh oh!

benadamstyles commented Dec 16, 2025

Uh oh!

Choose a reason for hiding this comment

illusionalsagacity commented Dec 2, 2025 •

edited

Loading

illusionalsagacity commented Dec 3, 2025 •

edited

Loading