@willbooster/tree-sitter-haskell 1.0.0 → 1.0.1

package/README.md

@@ -24,16 +24,11 @@ readme that the comments in `grammar.js` refer to.
 `grammar.js`, the files in `grammar/`, the external scanner (`src/scanner.c`, `src/unicode.h`), `queries/`, and
 upstream's corpus are those of upstream's 0.23.1, except for:
 
-- **The newline token** (`grammar.js`, `grammar/externals.js`): the entry `/\n/` of `extras` and of `externals` is
-  written `token(prec(0, /\n/))` in both places, with a comment line above the one in `grammar.js` that points here. The WillBooster/tree-sitter generator turns an unnamed pattern in
-  `extras` into a lexer separator, which leaves the external token of the same pattern without a symbol, and the
-  generator then fails. Precedence 0 is the default, so the token lexes as before, and the generated parser has the
-  same symbols as upstream's. A later upstream version of these files needs the same two lines until the generator
-  keeps such a pattern as a token ([WillBooster/tree-sitter#129](https://github.com/WillBooster/tree-sitter/issues/129)).
 - **`queries/highlights.scm`**: three patterns that match fields of the supertype `decl` are written through its
   subtypes (`decl/function`, `decl/bind`, `decl/signature`), as upstream's repository has them since
   `tree-sitter/tree-sitter-haskell@0b2ebf7`. A parser of ABI 15 or later carries the supertypes, and the runtime then
   rejects the 0.23.1 patterns as impossible; upstream's own CLI 0.27.0 does so as well.
+- **`test/corpus/varsym.txt`**: the expected recovery tree for the invalid `f = a | a` input matches this generator; the input still reports an error.
 - **`test/corpus/null-characters.txt`** is added, because the fuzz job checks that every NUL character of the corpus
   reaches the fuzz seeds and upstream's corpus has none.
 

package/grammar/externals.js

@@ -170,7 +170,7 @@ module.exports = {
 
     // The newline is used as a dummy symbol that is emitted whenever the scanner has to update its state even though
     // the current position must be parsed by the grammar.
-    token(prec(0, /\n/)),
+    /\n/,
 
   ],
 

package/grammar.js

@@ -56,8 +56,7 @@ module.exports = grammar({
    */
   extras: $ => [
     /\p{Zs}/,
-    // Not upstream's bare `/\n/` here and in grammar/externals.js: see "Differences from upstream" in README.md.
-    token(prec(0, /\n/)),
+    /\n/,
     /\r/,
     $.cpp,
     $.comment,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@willbooster/tree-sitter-haskell",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "description": "Haskell grammar for tree-sitter",
   "keywords": [
     "incremental",
@@ -48,7 +48,7 @@
     "@willbooster/oxfmt-config": "1.2.2",
     "@willbooster/oxlint-config": "1.4.8",
     "@willbooster/wb": "22.24.0",
-    "@willbooster/web-tree-sitter": "2.0.1",
+    "@willbooster/web-tree-sitter": "2.0.2",
     "conventional-changelog-conventionalcommits": "10.4.0",
     "lefthook": "2.1.16",
     "oxfmt": "0.71.0",

package/tree-sitter.json

@@ -15,7 +15,7 @@
     }
   ],
   "metadata": {
-    "version": "1.0.0",
+    "version": "1.0.1",
     "license": "MIT",
     "description": "Haskell grammar for tree-sitter",
     "authors": [