Add linkifier rule to inline chain for full links

prevents emphasis from appearing in `http://example.org/foo._bar_.baz`

CHANGELOG.md

@@ -19,6 +19,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ### Fixed
 - Smartquotes, typographic replacements and plain text links can now be escaped
   with backslash (e.g. `\(c)` or `google\.com` are no longer replaced).
+- Fixed collision of emphasis and linkifier (so `http://example.org/foo._bar_-_baz`
+  is now a single link, not emphasized). Emails and fuzzy links are not affected by this.
 
 
 ## [12.3.2] - 2022-01-08

lib/parser_inline.js

@@ -14,6 +14,7 @@ var Ruler           = require('./ruler');
 
 var _rules = [
   [ 'text',            require('./rules_inline/text') ],
+  [ 'linkify',         require('./rules_inline/linkify') ],
   [ 'newline',         require('./rules_inline/newline') ],
   [ 'escape',          require('./rules_inline/escape') ],
   [ 'backticks',       require('./rules_inline/backticks') ],

lib/rules_core/linkify.js

@@ -69,8 +69,17 @@ module.exports = function linkify(state) {
         level = currentToken.level;
         lastPos = 0;
 
-        for (ln = 0; ln < links.length; ln++) {
+        // forbid escape sequence at the start of the string,
+        // this avoids http\://example.com/ from being linkified as
+        // http:<a href="//example.com/">//example.com/</a>
+        if (links.length > 0 &&
+            links[0].index === 0 &&
+            i > 0 &&
+            tokens[i - 1].type === 'text_special') {
+          links = links.slice(1);
+        }
 
+        for (ln = 0; ln < links.length; ln++) {
           url = links[ln].url;
           fullUrl = state.md.normalizeLink(url);
           if (!state.md.validateLink(fullUrl)) { continue; }

lib/rules_inline/html_inline.js

@@ -6,6 +6,14 @@
 var HTML_TAG_RE = require('../common/html_re').HTML_TAG_RE;
 
 
+function isLinkOpen(str) {
+  return /^<a[>\s]/i.test(str);
+}
+function isLinkClose(str) {
+  return /^<\/a\s*>/i.test(str);
+}
+
+
 function isLetter(ch) {
   /*eslint no-bitwise:0*/
   var lc = ch | 0x20; // to lower case
@@ -41,6 +49,9 @@ module.exports = function html_inline(state, silent) {
   if (!silent) {
     token         = state.push('html_inline', '', 0);
     token.content = state.src.slice(pos, pos + match[0].length);
+
+    if (isLinkOpen(token.content))  state.linkLevel++;
+    if (isLinkClose(token.content)) state.linkLevel--;
   }
   state.pos += match[0].length;
   return true;

lib/rules_inline/link.js

@@ -137,7 +137,9 @@ module.exports = function link(state, silent) {
       attrs.push([ 'title', title ]);
     }
 
+    state.linkLevel++;
     state.md.inline.tokenize(state);
+    state.linkLevel--;
 
     token        = state.push('link_close', 'a', -1);
   }

lib/rules_inline/linkify.js

@@ -0,0 +1,58 @@
+// Process links like https://example.org/
+
+'use strict';
+
+
+// RFC3986: scheme = ALPHA *( ALPHA / DIGIT / "+" / "-" / "." )
+var SCHEME_RE = /(?:^|[^a-z0-9.+-])([a-z][a-z0-9.+-]*)$/i;
+
+
+module.exports = function linkify(state, silent) {
+  var pos, max, match, proto, link, url, fullUrl, token;
+
+  if (!state.md.options.linkify) return false;
+  if (state.linkLevel > 0) return false;
+
+  pos = state.pos;
+  max = state.posMax;
+
+  if (pos + 3 > max) return false;
+  if (state.src.charCodeAt(pos) !== 0x3A/* : */) return false;
+  if (state.src.charCodeAt(pos + 1) !== 0x2F/* / */) return false;
+  if (state.src.charCodeAt(pos + 2) !== 0x2F/* / */) return false;
+
+  match = state.pending.match(SCHEME_RE);
+  if (!match) return false;
+
+  proto = match[1];
+
+  link = state.md.linkify.matchAtStart(state.src.slice(pos - proto.length));
+  if (!link) return false;
+
+  url = link.url;
+
+  // disallow '*' at the end of the link (conflicts with emphasis)
+  url = url.replace(/\*+$/, '');
+
+  fullUrl = state.md.normalizeLink(url);
+  if (!state.md.validateLink(fullUrl)) return false;
+
+  if (!silent) {
+    state.pending = state.pending.slice(0, -proto.length);
+
+    token         = state.push('link_open', 'a', 1);
+    token.attrs   = [ [ 'href', fullUrl ] ];
+    token.markup  = 'linkify';
+    token.info    = 'auto';
+
+    token         = state.push('text', '', 0);
+    token.content = state.md.normalizeLinkText(url);
+
+    token         = state.push('link_close', 'a', -1);
+    token.markup  = 'linkify';
+    token.info    = 'auto';
+  }
+
+  state.pos += url.length - proto.length;
+  return true;
+};

lib/rules_inline/state_inline.js

@@ -35,6 +35,10 @@ function StateInline(src, md, env, outTokens) {
   // backtick length => last seen position
   this.backticks = {};
   this.backticksScanned = false;
+
+  // Counter used to disable inline linkify-it execution
+  // inside <a> and markdown links
+  this.linkLevel = 0;
 }
 
 

package.json

@@ -39,7 +39,7 @@
   "dependencies": {
     "argparse": "^2.0.1",
     "entities": "~3.0.1",
-    "linkify-it": "^3.0.1",
+    "linkify-it": "markdown-it/linkify-it",
     "mdurl": "^1.0.1",
     "uc.micro": "^1.0.5"
   },

test/fixtures/markdown-it/linkify.txt

@@ -30,6 +30,86 @@ don't touch text in html <a> tags
 .
 
 
+entities inside raw links
+.
+https://example.com/foo&amp;bar
+.
+<p><a href="https://example.com/foo&amp;amp;bar">https://example.com/foo&amp;amp;bar</a></p>
+.
+
+
+emphasis inside raw links (asterisk, can happen in links with params)
+.
+https://example.com/foo*bar*baz
+.
+<p><a href="https://example.com/foo*bar*baz">https://example.com/foo*bar*baz</a></p>
+.
+
+
+emphasis inside raw links (underscore)
+.
+http://example.org/foo._bar_-_baz
+.
+<p><a href="http://example.org/foo._bar_-_baz">http://example.org/foo._bar_-_baz</a></p>
+.
+
+
+backticks inside raw links
+.
+https://example.com/foo`bar`baz
+.
+<p><a href="https://example.com/foo%60bar%60baz">https://example.com/foo`bar`baz</a></p>
+.
+
+
+links inside raw links
+.
+https://example.com/foo[123](456)bar
+.
+<p><a href="https://example.com/foo%5B123%5D(456)bar">https://example.com/foo[123](456)bar</a></p>
+.
+
+
+escapes not allowed at the start
+.
+\https://example.com
+.
+<p>\https://example.com</p>
+.
+
+
+escapes not allowed at comma
+.
+https\://example.com
+.
+<p>https://example.com</p>
+.
+
+
+escapes not allowed at slashes
+.
+https:\//aa.org https://bb.org
+.
+<p>https://aa.org <a href="https://bb.org">https://bb.org</a></p>
+.
+
+
+fuzzy link shouldn't match cc.org
+.
+https:/\/cc.org
+.
+<p>https://cc.org</p>
+.
+
+
+bold links (exclude markup of pairs from link tail)
+.
+**http://example.com/foobar**
+.
+<p><strong><a href="http://example.com/foobar">http://example.com/foobar</a></strong></p>
+.
+
+
 match links without protocol
 .
 www.example.org
@@ -55,3 +135,35 @@ http://example.com/(c)
 .
 <p><a href="http://example.com/(c)">http://example.com/(c)</a></p>
 .
+
+
+coverage, prefix not valid
+.
+http:/example.com/
+.
+<p>http:/example.com/</p>
+.
+
+
+coverage, negative link level
+.
+</a>[https://example.com](https://example.com)
+.
+<p></a><a href="https://example.com"><a href="https://example.com">https://example.com</a></a></p>
+.
+
+
+emphasis with '*', real link:
+.
+http://cdecl.ridiculousfish.com/?q=int+%28*f%29+%28float+*%29%3B
+.
+<p><a href="http://cdecl.ridiculousfish.com/?q=int+%28*f%29+%28float+*%29%3B">http://cdecl.ridiculousfish.com/?q=int+(*f)+(float+*)%3B</a></p>
+.
+
+
+emphasis with '_', real link:
+.
+https://www.sell.fi/sites/default/files/elainlaakarilehti/tieteelliset_artikkelit/kahkonen_t._et_al.canine_pancreatitis-_review.pdf
+.
+<p><a href="https://www.sell.fi/sites/default/files/elainlaakarilehti/tieteelliset_artikkelit/kahkonen_t._et_al.canine_pancreatitis-_review.pdf">https://www.sell.fi/sites/default/files/elainlaakarilehti/tieteelliset_artikkelit/kahkonen_t._et_al.canine_pancreatitis-_review.pdf</a></p>
+.