Code folding #6
Code folding #6
Conversation
Should fold the following code correctly
```tsx
// Hey Emacs, this is -*- coding: utf-8 -*-
'use client';
import { Provider, defaultTheme } from '@adobe/react-spectrum';
interface Props {
children: React.ReactNode;
}
export const App = ({ children }: Props): JSX.Element => {
return (
<Provider
theme={defaultTheme}
// locale={yourLocaleHere}
>
<main className="flex min-h-screen flex-col items-center justify-between p-24">
{children}
</main>
<div>
test
</div>
</Provider>
);
};
```
Emacs 29 has forward-sexp (&optional arg interactive) arguments rather than (forward-sexp n) The calls to forward-sexp() should be compatible between Emacs versions as long as INTERACTIVE argument is not used.
Should fold 'content' value correctly in the following snippet:
```javascript
module.exports = {
important: '#app',
content: [
'./src/pages/**/*.{js,ts,jsx,tsx,mdx}',
'./src/components/**/*.{js,ts,jsx,tsx,mdx}',
'./src/app/**/*.{js,ts,jsx,tsx,mdx}',
],
presets: [spectrumPreset],
plugins: [],
};
```
... and guarding against recursive call when forward-sexp-function variable points to jtsx-hs-forward-sexp function
...in jtsx modes that handle JSX/TSX blocks and some minor cleanups
Limitations: * jtsx-backward-up-list() does not support backward-up-list arguments - always jumps one level up. Also, need to investigate further if such function is required. It might be possible to achive similar functionality with plain backward-up-list() by refactoring jtsx-hs-forward-sexp() behaviour.
|
Thanks, I will review it as soon I have some free time ! Pasting here the description of the PR done in #5.
|
There was a problem hiding this comment.
Just finished to review your PR.
Here the summary of what I did:
- Test your changes -> OK for me !
- Add 2 commits : minor changes in the documentation and some new tests
- Leave 2 comments in the review
I have 2 more questions/remarks :
- It appears that
jtsx-hs-forward-sexpis not only intented forhideshownow, maybe we should rename it intojtsx-forward-sexp? jtsx-backward-up-listis independant ofhideshowand is intended to be used as a replacement ofbackward-up-listforjtsxmodes, if I am not wrong ? So we should document it as a public function in the README I guess (and warning about restrictions) ?
Concerning your concerns:
The hide-show code is a blend of treesit and regexp approaches. It’s uncertain if this is the best practice. It might need to be refactored to rely solely on treesit queries.
Yes but I don't think there any issues with that excepted in terms of performance.
For having a pure tree-sitter code folding, we should use something like ts-fold. However it is not on a package manager yet.
hide-show does not fold js/ts code inside {} fragments in jsx blocks – it only folds the entire {} fragment.
The function jtsx-hs-forward-sexp does not handle js/ts blocks inside jsx blocks.
js code nested in jsx is not straightforward to detect with tree-sitter. I worked on that point few weeks ago for the comment/uncomment function. But it is a quite hacky implementation. I think it would need to be rewritten to be easily used outside of comment/uncomment functionality.
The function jtsx-hs-forward-sexp does not fully support forward-sexp arguments - it always jumps one sexp forward for a positive argument and one sexp backward for a negative argument.
The function jtsx-backward-up-list does not support backward-up-list arguments - it always jumps up one level.
Ok. Anyway, it is a good start, and it could be enhanced later if necessary.
|
Thank you for merging. Great work! In the upcoming iteration, I plan to delve into |
|
Glad to hear you are planning some others contributions.
I am just remembering that
I have activated |
Some code folding fixes
see #5 for details