Fix trimming of whitespace around breaks

Related-to: remarkjs/remark#1202.
syntax-tree · Aug 9, 2023 · 7ff28fb · 7ff28fb
1 parent b4ec6fa
commit 7ff28fb
Show file tree

Hide file tree

Showing 3 changed files with 55 additions and 2 deletions.
diff --git a/lib/state.js b/lib/state.js
@@ -291,14 +291,14 @@ export function createState(tree, options) {
         if (result) {
           if (index && nodes[index - 1].type === 'break') {
             if (!Array.isArray(result) && result.type === 'text') {
-              result.value = result.value.replace(/^\s+/, '')
+              result.value = trimMarkdownSpaceStart(result.value)
             }
 
             if (!Array.isArray(result) && result.type === 'element') {
               const head = result.children[0]
 
               if (head && head.type === 'text') {
-                head.value = head.value.replace(/^\s+/, '')
+                head.value = trimMarkdownSpaceStart(head.value)
               }
             }
           }
@@ -447,3 +447,23 @@ export function wrap(nodes, loose) {
 
   return result
 }
+
+/**
+ * Trim spaces and tabs at the start of `value`.
+ *
+ * @param {string} value
+ *   Value to trim.
+ * @returns {string}
+ *   Result.
+ */
+function trimMarkdownSpaceStart(value) {
+  let index = 0
+  let code = value.charCodeAt(index)
+
+  while (code === 9 || code === 32) {
+    index++
+    code = value.charCodeAt(index)
+  }
+
+  return value.slice(index)
+}
diff --git a/package.json b/package.json
@@ -109,6 +109,7 @@
       "import/no-cycle": "error",
       "max-depth": "off",
       "unicorn/prefer-at": "off",
+      "unicorn/prefer-code-point": "off",
       "unicorn/prefer-string-replace-all": "off"
     }
   }

diff --git a/test/break.js b/test/break.js
@@ -45,4 +45,36 @@ test('break', async function (t) {
       h('p', ['alpha', h('br'), '\n', h('em', 'bravo')])
     )
   })
+
+  await t.test('should trim text after a `br` (#3)', async function () {
+    assert.deepEqual(
+      toHast({
+        type: 'paragraph',
+        children: [
+          {type: 'text', value: 'a'},
+          {type: 'break'},
+          // U+3000 (ideographic space).
+          {type: 'text', value: '　b'},
+          {type: 'break'},
+          // U+2003 (em space).
+          {type: 'text', value: ' c'},
+          {type: 'break'},
+          // U+00A0 (no-break space).
+          {type: 'text', value: ' d'}
+        ]
+      }),
+      h('p', [
+        'a',
+        h('br'),
+        '\n',
+        '　b',
+        h('br'),
+        '\n',
+        ' c',
+        h('br'),
+        '\n',
+        ' d'
+      ])
+    )
+  })
 })