fix: allow zero-width space as emphasis separator (#112)

fixes #98.

.changeset/thick-mice-breathe.md

@@ -0,0 +1,5 @@
+---
+'uniorg-parse': patch
+---
+
+Allow Zero-Width Space as emphasis separator.

packages/uniorg-parse/src/__snapshots__/parser.spec.ts.snap

@@ -5343,3 +5343,20 @@ children:
           minute: 36
         end: null
 `;
+
+exports[`org/parser zero-width whitespace as emphasis separator 1`] = `
+type: "org-data"
+contentsBegin: 0
+contentsEnd: 12
+children:
+  - type: "paragraph"
+    affiliated: {}
+    contentsBegin: 0
+    contentsEnd: 12
+    children:
+      - type: "code"
+        value: "code"
+        children: []
+      - type: "text"
+        value: "​hello"
+`;

packages/uniorg-parse/src/parse-options.ts

@@ -38,14 +38,19 @@ export interface ParseOptions {
 export const defaultOptions: ParseOptions = {
   todoKeywords: ['TODO', 'DONE'],
   useSubSuperscripts: true,
+  // Interestingly enough, zero-width space (\u200b) is not considered
+  // a space in unicode but is considered a space by Emacs. This is
+  // why we have to add \u200b explicitly after \s in the
+  // regex. Otherwise, the suggested use-case of adding ZWSP as a
+  // markup border does not work.
   emphasisRegexpComponents: {
     // deviates from org mode default to allow ndash, mdash, and
     // quotes (’“”)
-    pre: '-–—\\s\\(\'’"“”\\{',
+    pre: '-–—\\s\u200b\\(\'’"“”\\{',
     // deviates from org mode default to allow ndash, mdash, and
     // quotes (’“”)
-    post: '-–—\\s.,:!?;\'’"“”\\)\\}\\[',
-    border: '\\s',
+    post: '-–—\\s\u200b.,:!?;\'’"“”\\)\\}\\[',
+    border: '\\s\u200b',
     body: '.',
     newline: 1,
   },

packages/uniorg-parse/src/parser.spec.ts

@@ -1127,4 +1127,7 @@ line break`
 not next line`
     );
   });
+
+  // See https://github.com/rasendubi/uniorg/issues/98
+  itParses('zero-width whitespace as emphasis separator', `~code~\u200bhello`);
 });