steid

feat: a .jsx file reads as JavaScript rather than as plain text

two-face's JavaScript grammar registers `js` and `htc` and nothing else, so `.jsx`
came out colourless while `.tsx` was highlighted — TypeScriptReact is its own
grammar and JavaScriptReact is not. JSX is JavaScript with element literals in it
and the JavaScript grammar reads it well enough.

An `ALIASES` table rather than a special case, consulted *after* the syntax set has
had its say: an alias is what to do when nothing claims an extension, never a way
to overrule a grammar that does. One entry, and it should stay near one — a wrong
alias reads worse than no colour.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01J18ViwAfdswUCMb2DXJZFG

JamesPatrickGill authored 14 hours agoparent00d9c06Browse files610ab384a91b2fd66840bcff0aa636c69ee5d8ef

2 files changed+27 −3

plans/current.md+3 −2View file

		@@ -142,8 +142,9 @@ instance, and Steid's own source is pushed to it and browsable there.
142	142	- Diffs, READMEs and markdown code fences are unhighlighted. The README goes through
143	143	`web/markdown.rs`, which writes its own `<pre><code>`; wiring `highlight.rs` into it is
144	144	a small follow-up.
145		−- `.jsx` is plain: two-face's JavaScript grammar claims only `js`/`htc`; `tsx` is
146		− covered by TypeScriptReact. An alias is one line.
	145	+- ~~`.jsx` is plain~~ — aliased onto JavaScript. `highlight.rs` gained an `ALIASES`
	146	+ table, consulted only after the syntax set's own answer, so an alias can never
	147	+ overrule a real grammar. It has one entry and should stay small.
147	148	- `\r\n` files keep the `\r` inside highlighted markup. Invisible under
148	149	`white-space: pre`; noted rather than fixed.
149	150	- The blob header does not name the detected language. It is only visible as colour.

src/infrastructure/highlight.rs+24 −1View file

		@@ -121,6 +121,17 @@ fn syntaxes() -> &'static SyntaxSet {
121	121	SYNTAXES.get_or_init(two_face::syntax::extra_newlines)
122	122	}
123	123
	124	+/// Extensions no grammar in the set claims, and the grammar to read them with anyway.
	125	+///
	126	+/// `two-face`'s JavaScript grammar registers `js` and `htc` and nothing else, so `.jsx`
	127	+/// rendered as plain text while `.tsx` was highlighted, TypeScriptReact being its own
	128	+/// grammar. JSX is JavaScript with element literals in it and the JavaScript grammar
	129	+/// reads it well enough.
	130	+///
	131	+/// One entry, deliberately: every alias is a claim that two languages are close enough
	132	+/// to read as one, and a wrong one is worse than no colour.
	133	+const ALIASES: &[(&str, &str)] = &[("jsx", "js")];
	134	+
124	135	/// The language, from the file name and then the first line.
125	136	///
126	137	/// This is `SyntaxSet::find_syntax_for_file` without the filesystem: that method opens
		@@ -135,9 +146,19 @@ fn syntax_for<'a>(
135	146	) -> Option<&'a SyntaxReference> {
136	147	let extension = file_name.rsplit_once('.').map(\|(_, ext)\| ext);
137	148
	149	+ // After the set's own answer, never before it: an alias is what to do when no
	150	+ // grammar claims the extension, not a way to overrule one that does.
	151	+ let aliased = extension.and_then(\|ext\| {
	152	+ ALIASES
	153	+ .iter()
	154	+ .find(\|(from, _)\| *from == ext)
	155	+ .map(\|(_, to)\| *to)
	156	+ });
	157	+
138	158	let syntax = syntaxes
139	159	.find_syntax_by_extension(file_name)
140	160	.or_else(\|\| extension.and_then(\|ext\| syntaxes.find_syntax_by_extension(ext)))
	161	+ .or_else(\|\| aliased.and_then(\|to\| syntaxes.find_syntax_by_extension(to)))
141	162	.or_else(\|\| syntaxes.find_syntax_by_first_line(text.lines().next().unwrap_or("")))?;
142	163
143	164	(syntax.name != syntaxes.find_syntax_plain_text().name).then_some(syntax)
		@@ -310,7 +331,8 @@ mod tests {
310	331
311	332	/// The gap `two_face` was added to close, asserted rather than described: TOML,
312	333	/// Dockerfile and TypeScript rendered plain until it landed, and the first two are
313		− /// in this repository. `.jsx` is still absent — see the handover.
	334	+ /// in this repository. `.jsx` is here through [`ALIASES`] rather than through a
	335	+ /// grammar of its own.
314	336	#[test]
315	337	fn the_languages_this_repository_uses_are_all_highlighted() {
316	338	for name in [
		@@ -327,6 +349,7 @@ mod tests {
327	349	"Makefile",
328	350	"app.py",
329	351	"index.js",
	352	+ "App.jsx",
330	353	"main.go",
331	354	"query.sql",
332	355	"main.c",