Remove support of children spread inside JSX

[Copilot]

This PR removes support for JSX children spread syntax while preserving props spread functionality.

What was removed

Children spread syntax like this is no longer supported:

<div> ...children </div>
<Component> ...rest </Component>
<> ...items </>

What still works

Props spread continues to work as before:

<Button {...buttonProps} />
<div className="wrapper" {...divProps} />
<MyComponent id="test" {...rest} onClick={handler} />

Normal JSX children also continue to work:

<div> {text} {otherElement} </div>
<Container> <span>child</span> </Container>

Implementation details

The changes remove the JSXChildrenSpreading variant from the jsx_children type and update all related code:

• Parser: Updated parse_jsx_children to no longer recognize ... syntax in children context
• Type system: Removed JSXChildrenSpreading of expression from jsx_children type
• Processing: Updated all AST mappers, iterators, printers, and analysis tools
• Tests: Removed children spread examples from test files

The implementation is minimal and surgical - only the children spread functionality is removed while all other JSX features remain unchanged.

Files changed

• Core type definition in parsetree.ml
• Parser logic in res_core.ml and printer in res_printer.ml
• AST processing across 8 files in compiler/ml/
• Syntax processing in 3 files in compiler/syntax/src/
• Frontend and analysis support files
• Test cases updated to remove children spread examples

This change makes the JSX syntax more consistent by having spread syntax only available in props context.

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[pkg-pr-new]

Open in StackBlitz

rescript

npm i https://pkg.pr.new/rescript-lang/rescript@7869

@rescript/darwin-arm64

npm i https://pkg.pr.new/rescript-lang/rescript/@rescript/darwin-arm64@7869

@rescript/darwin-x64

npm i https://pkg.pr.new/rescript-lang/rescript/@rescript/darwin-x64@7869

@rescript/linux-arm64

npm i https://pkg.pr.new/rescript-lang/rescript/@rescript/linux-arm64@7869

@rescript/linux-x64

npm i https://pkg.pr.new/rescript-lang/rescript/@rescript/linux-x64@7869

@rescript/runtime

npm i https://pkg.pr.new/rescript-lang/rescript/@rescript/runtime@7869

@rescript/win32-x64

npm i https://pkg.pr.new/rescript-lang/rescript/@rescript/win32-x64@7869

commit: 605b78d

[nojaf]

Looks good at a glance. Can we add a test to check which parsing error we now get for ...? If that’s too cryptic, I’d raise a custom parser error that clearly states child spreading was removed.

[zth]

This can be removed now as well, right? Should probably do that in the same PR.

[tsnobip]

fixed!

[tsnobip]

@zth let me know if you additional comments!

[tsnobip]

Looks good at a glance. Can we add a test to check which parsing error we now get for ...? If that’s too cryptic, I’d raise a custom parser error that clearly states child spreading was removed.

@nojaf that's a good point, unfortunately the error is indeed quite cryptic!

  Syntax error!
  syntax_tests/data/parsing/errors/expressions/jsx.res:18:15-17

  16 │ 
  17 │ // spread children
  18 │ let x = <div> ...c </div>

  Did you forget a `</` here?


  Syntax error!
  syntax_tests/data/parsing/errors/expressions/jsx.res:18:15-17

  16 │ 
  17 │ // spread children
  18 │ let x = <div> ...c </div>

  I'm not sure what to parse here when looking at "...".


  Syntax error!
  syntax_tests/data/parsing/errors/expressions/jsx.res:18:22-25

  16 │ 
  17 │ // spread children
  18 │ let x = <div> ...c </div>

  unterminated regex

[nojaf]

@tsnobip inside the loop you probably want to match ... and give a better parse error.

[tsnobip]

@nojaf is it better like that?

[nojaf]

Yes, does wrapping it {...c} do anything?
Syntax-wise probably that is going to be valid.
Would surprise me if that type-checks though.

[tsnobip]

@nojaf no, it would not type-checks.
But that's actually why I wanted first to merge this PR, wrapping will do ...{c} and the compiler will complain that spreading children is no longer supported!

[nojaf]

Great work!

[glennsl]

Why was this removed? And why is the only description of this change a useless AI summary that says nothing about the rationale or possible consequences of such a change? Have we completely abandoned any attempt at human-human interaction now?

[nojaf]

If I recall correctly, the use-case was very limited that it didn't seem worth keeping.

https://rescript-lang.org/docs/manual/v9.0.0/jsx#children-spread

The output in v11 seems to be the same: https://rescript-lang.org/try?version=v11.1.4&module=esmodule&code=DYUwLgBAnhC8EB4BGAnCB6AfAKFJAXnIgCYCWAbptIumZbuBAB5EJ1UB0XMCtFOQA so that lead us to believe this was no longer relevant.

Were you using this?

[glennsl]

Thanks @nojaf. Yes, I've been using it quite extensively, But it does indeed seem to be unnecessary now, which makes this change a lot less invasive than I thought it was.

[glennsl]

And sorry for lashing out. I have severe AI fatigue and am seriously considering quitting software development altogether because of it.

[nojaf]

No worries, I can relate to that 😊.

Another reason to drop this, is that we have some post v12 ideas to bring JSX syntax closer to the JavaScript JSX. And not having to deal with this would benefit that.