allow step attributes to be valueless

package.json

@@ -1,6 +1,6 @@
 {
   "name": "ecmarkdown",
-  "version": "8.0.0",
+  "version": "8.1.0",
   "description": "A compiler for \"Ecmarkdown\" algorithm shorthand into HTML.",
   "main": "dist/ecmarkdown.js",
   "scripts": {

src/tokenizer.ts

@@ -3,7 +3,7 @@ import type { Unlocated, Token, AttrToken, Position } from './node-types';
 const fieldOrSlotRegexp = /^\[\[[a-zA-Z0-9_]+\]\]/;
 const tagRegexp = /^<[/!]?(\w[\w-]*)(\s+\w[\w-]*(\s*=\s*("[^"]*"|'[^']*'|[^><"'=`]+))?)*\s*>/;
 const commentRegexp = /^<!--[\w\W]*?-->/;
-const attrRegexp = /^\[ *[\w-]+ *= *"(?:[^"\\\x00-\x1F]|\\["\\/bfnrt]|\\u[a-fA-F]{4})*" *(?:, *[\w-]+ *= *"(?:[^"\\\x00-\x1F]|\\["\\/bfnrt]|\\u[a-fA-F]{4})*" *)*] /;
+const attrRegexp = /^\[ *[\w-]+ *(?:= *"(?:[^"\\\x00-\x1F]|\\["\\/bfnrt]|\\u[a-fA-F]{4})*")? *(?:, *[\w-]+ *(?:= *"(?:[^"\\\x00-\x1F]|\\["\\/bfnrt]|\\u[a-fA-F]{4})*")? *)*] /;
 const digitRegexp = /\d/;
 
 const opaqueTags = new Set(['emu-grammar', 'emu-production', 'pre', 'code', 'script', 'style']);
@@ -158,13 +158,17 @@ export class Tokenizer {
 
   // attribute tokens are only valid immediately after list tokens, so we let this be called by the parser.
   tryScanListItemAttributes() {
-    const match = this.str.slice(this.pos).match(attrRegexp);
+    const rest = this.str.slice(this.pos);
+    const match = rest.match(attrRegexp);
     if (!match) {
+      if (rest.startsWith('[')) {
+        this.raise('could not parse attributes for step', this.getLocation());
+      }
       return [];
     }
 
     const parts = match[0].matchAll(
-      /([\w-]+) *= *("(?:[^"\\\x00-\x1F]|\\["\\/bfnrt]|\\u[a-fA-F]{4})*")/g
+      /([\w-]+) *(?:= *("(?:[^"\\\x00-\x1F]|\\["\\/bfnrt]|\\u[a-fA-F]{4})*"))?/g
     );
     const tokens = [];
     let offset = 0;
@@ -178,7 +182,7 @@ export class Tokenizer {
       const tok: Unlocated<AttrToken> = {
         name: 'attr',
         key,
-        value: JSON.parse(value),
+        value: value == null ? '' : JSON.parse(value), // the parse is guaranteed to succeed because the regex is quite restricted
       };
       this.pos += part.length;
       this.locate(tok, tokStart);

test/cases/list-attrs.ecmarkdown

@@ -2,6 +2,7 @@
 1. Item
 1. Item
   1. [attr="bar"] Item
+  1. [no-value, with-empty="", with-value="val"] Item
   1. Item
 1. Item
   * Unordered item

test/cases/list-attrs.html

@@ -3,6 +3,7 @@
   <li>Item</li>
   <li>Item<ol>
       <li attr="bar">Item</li>
+      <li no-value with-empty with-value="val">Item</li>
       <li>Item</li>
     </ol>
   </li>

test/errors.js

@@ -68,4 +68,13 @@ ${M}      * b
       'Unexpected token parabreak; expected EOF'
     );
   });
+
+  it('bad step attributes', function () {
+    assertAlgError(
+      positioned`
+      1. ${M}[x y]
+    `,
+      'could not parse attributes for step'
+    );
+  });
 });

test/parser.js

@@ -4,14 +4,16 @@ const { parseAlgorithm, parseFragment } = require('../');
 
 describe('Parser', function () {
   it('tracks positions in algorithms', function () {
-    const baseSource = '  1. [id="thing"] a\n  2. b c\n    <figure>text</figure>';
+    const baseSource = '  1. [id="thing", other] a\n  2. b c\n    <figure>text</figure>';
     const assertNodeLocation = makeAssertLocation(baseSource);
     const algorithm = parseAlgorithm(baseSource);
     assertNodeLocation(algorithm, baseSource);
     const list = algorithm.contents;
-    assertNodeLocation(list, '  1. [id="thing"] a\n  2. b c\n    <figure>text</figure>');
+    assertNodeLocation(list, '  1. [id="thing", other] a\n  2. b c\n    <figure>text</figure>');
     const item0 = list.contents[0];
-    assertNodeLocation(item0, '  1. [id="thing"] a\n');
+    assertNodeLocation(item0, '  1. [id="thing", other] a\n');
+    assertNodeLocation(item0.attrs[0], 'id="thing"');
+    assertNodeLocation(item0.attrs[1], 'other');
     assertNodeLocation(item0.contents[0], 'a');
     const item1 = list.contents[1];
     assertNodeLocation(item1, '  2. b c\n    <figure>text</figure>');