Zulip server and web application. Open-source team chat that helps teams stay productive and focused.
marked: Keep balanced parentheses in link URLs.
Shubham Padiad7c57ed15b82
diff --git a/web/third/marked/lib/marked.cjs b/web/third/marked/lib/marked.cjsindex 973f383808f..c5654ae01bb 100644--- a/web/third/marked/lib/marked.cjs+++ b/web/third/marked/lib/marked.cjs@@ -718,6 +718,7 @@ InlineLexer.prototype.output = function(src) {// linkif (cap = this.rules.link.exec(src)) {+ cap = this.reparseLinkHref(src, cap);src = src.substring(cap[0].length);this.inLink = true;out += this.outputLink(cap, {@@ -867,6 +868,68 @@ InlineLexer.prototype.output = function(src) {* Compile Link*/+// An optional quoted title, then the ")" that closes the link:+// ) or "some title")+var linkEnding = /^(?:\s+(['"])([\s\S]*?)\1)?\s*\)/;++// The link regex stops the href at the first ")". That is wrong when+// the URL itself contains parentheses, like http://a/b(c)d: the href+// is cut at "c" and "d)" is left over as text.+//+// `cap` is the regex's match:+// cap[0] the text the regex matched, from "[" to ")"+// cap[1] the link text+// cap[2] the href+// cap[3] the title, or undefined+// This returns an array with the same four entries. cap[0] now ends+// at the ")" that really closes the link, and cap[2] holds the full+// href. cap[1] is copied over unchanged. If the link doesn't look the+// way we expect, `cap` is returned as is.+//+// This follows the server's markdown implementation. An href wrapped+// in <> runs up to the ">". Otherwise, a quoted string followed by+// ")" is the title and the href runs up to it; the title may itself+// contain parentheses. With no title, the href runs up to the ")"+// that balances the link's "(", even across spaces.+InlineLexer.prototype.reparseLinkHref = function(src, cap) {+ // Start just after "[text](", or " === '!' ? 2 : 1) + cap[1].length + 2;+ // Skip any whitespace before the href.+ index += /^\s*/.exec(src.substring(index))[0].length;+ var rest = src.substring(index);+ var end, ending;+ if (rest.charAt(0) === '<') {+ var close = rest.indexOf('>');+ if (close === -1) {+ return cap;+ }+ ending = linkEnding.exec(rest.substring(close + 1));+ if (!ending) {+ return cap;+ }+ index += close + 1 + ending[0].length;+ return [src.substring(0, index), cap[1], rest.substring(1, close), ending[2]];+ }+ // Find the ")" that has no matching "(" before it.+ var unmatched = findClosingBracket(rest, '()');+ // A quote after whitespace starts the title, as long as it comes+ // before that ")". Parentheses inside the title don't count.+ var quote = /\s['"]/.exec(rest);+ if (quote && (unmatched === -1 || quote.index < unmatched)) {+ ending = linkEnding.exec(rest.substring(quote.index));+ if (ending) {+ end = quote.index;+ index += quote.index + ending[0].length;+ return [src.substring(0, index), cap[1], rest.substring(0, end).replace(/\s+$/, ''), ending[2]];+ }+ }+ if (unmatched === -1) {+ return cap;+ }+ index += unmatched + 1;+ return [src.substring(0, index), cap[1], rest.substring(0, unmatched).replace(/\s+$/, ''), undefined];+};+InlineLexer.prototype.outputLink = function(cap, link) {var href = escape(link.href), title = link.title ? escape(link.title) : null;@@ -1387,6 +1450,27 @@ function unescape(html) {});}+// Adapted from marked v0.6.2 (https://github.com/markedjs/marked/pull/1476).+// Unlike upstream, backslash-escaped brackets are not skipped, to+// match the server's markdown implementation.+function findClosingBracket(str, b) {+ if (str.indexOf(b[1]) === -1) {+ return -1;+ }+ var level = 0;+ for (var i = 0; i < str.length; i++) {+ if (str[i] === b[0]) {+ level++;+ } else if (str[i] === b[1]) {+ level--;+ if (level < 0) {+ return i;+ }+ }+ }+ return -1;+}+function replace(regex, opt) {regex = regex.source;opt = opt || '';diff --git a/zerver/tests/fixtures/markdown_test_cases.json b/zerver/tests/fixtures/markdown_test_cases.jsonindex 44b9eec3700..4e8261ab464 100644--- a/zerver/tests/fixtures/markdown_test_cases.json+++ b/zerver/tests/fixtures/markdown_test_cases.json@@ -502,6 +502,94 @@"input": "[](<http://example.com/a b>)","expected_output": "<p><a href=\"http://example.com/a%20b\">http://example.com/a%20b</a></p>"},+ {+ "name": "link_with_parentheses_in_url",+ "input": "[this issue](https://github.com/zulip/zulip/issues?q=is%3Aissue%20state%3Aopen%20label%3A%22area%3A%20markdown%20(embeds%2Fpreviews)%22)",+ "expected_output": "<p><a href=\"https://github.com/zulip/zulip/issues?q=is%3Aissue%20state%3Aopen%20label%3A%22area%3A%20markdown%20(embeds%2Fpreviews)%22\">this issue</a></p>"+ },+ {+ "name": "link_with_nested_parentheses_in_url",+ "input": "[foo](http://example.com/a(b(c)d)e)",+ "expected_output": "<p><a href=\"http://example.com/a(b(c)d)e\">foo</a></p>"+ },+ {+ "name": "link_with_parentheses_in_url_and_title",+ "input": "[foo](http://example.com/a(b) \"my title\")",+ "expected_output": "<p><a href=\"http://example.com/a(b)\" title=\"my title\">foo</a></p>"+ },+ {+ "name": "link_with_unbalanced_parenthesis_in_title",+ "input": "[foo](http://example.com/a \"my title)\")",+ "expected_output": "<p><a href=\"http://example.com/a\" title=\"my title)\">foo</a></p>"+ },+ {+ "name": "link_with_parenthesis_in_url_before_title",+ "input": "[foo](http://example.com/a(b \"my title\") c)",+ "expected_output": "<p><a href=\"http://example.com/a(b\" title=\"my title\">foo</a> c)</p>"+ },+ {+ "name": "link_with_parenthesis_in_title",+ "input": "[foo](http://example.com/a \"my (title)\")",+ "expected_output": "<p><a href=\"http://example.com/a\" title=\"my (title)\">foo</a></p>"+ },+ {+ "name": "link_with_escaped_parenthesis_in_url",+ "input": "[foo](http://example.com/a\\)b)",+ "expected_output": "<p><a href=\"http://example.com/a\\\">foo</a>b)</p>"+ },+ {+ "name": "image_with_parentheses_in_url",+ "input": "c \"my title\")",+ "expected_output": "<p>c\">http://example.com/a(b)c</a> \"my title\")</p>",+ "marked_expected_output": "<p><img src=\"http://example.com/a(b)c\" alt=\"foo\" title=\"my title\"></p>"+ },+ {+ "name": "link_with_unbalanced_parenthesis_in_url",+ "input": "[foo](http://example.com/a(b)",+ "expected_output": "<p>[foo](<a href=\"http://example.com/a(b)\">http://example.com/a(b)</a></p>",+ "marked_expected_output": "<p><a href=\"http://example.com/a(b\">foo</a></p>"+ },+ {+ "name": "link_with_quote_and_parentheses_in_url",+ "input": "[film](https://en.wikipedia.org/wiki/Ocean's_Eleven_(2001_film))",+ "expected_output": "<p><a href=\"https://en.wikipedia.org/wiki/Ocean's_Eleven_(2001_film\">film</a>)</p>",+ "marked_expected_output": "<p><a href=\"https://en.wikipedia.org/wiki/Ocean's_Eleven_(2001_film)\">film</a></p>"+ },+ {+ "name": "link_with_parentheses_in_url_and_single_quoted_title",+ "input": "[foo](http://example.com/a(b)c 'my title')",+ "expected_output": "<p><a href=\"http://example.com/a(b)c\" title=\"my title\">foo</a></p>"+ },+ {+ "name": "link_with_parentheses_and_space_in_url",+ "input": "[foo](http://example.com/a(b c)d)",+ "expected_output": "<p><a href=\"http://example.com/a(b%20c)d\">foo</a></p>"+ },+ {+ "name": "link_with_space_in_url_and_title",+ "input": "[foo](http://example.com/a b \"my title\")",+ "expected_output": "<p><a href=\"http://example.com/a%20b\" title=\"my title\">foo</a></p>"+ },+ {+ "name": "link_with_quotes_in_url",+ "input": "[foo](http://example.com/a \"b\" c)",+ "expected_output": "<p><a href=\"http://example.com/a%20"b"%20c\">foo</a></p>"+ },+ {+ "name": "link_with_parentheses_in_angle_bracket_url",+ "input": "[foo](<http://example.com/a(b)c>)",+ "expected_output": "<p><a href=\"http://example.com/a(b)c\">foo</a></p>"+ },+ {+ "name": "link_with_parentheses_and_space_in_angle_bracket_url_and_title",+ "input": "[foo](<http://example.com/a(b c)d> \"my title\")",+ "expected_output": "<p><a href=\"http://example.com/a(b%20c)d\" title=\"my title\">foo</a></p>"+ },+ {+ "name": "link_with_unbalanced_parenthesis_in_angle_bracket_url",+ "input": "[foo](<http://example.com/a(b>)",+ "expected_output": "<p><a href=\"http://example.com/a(b\">foo</a></p>"+ },{"name": "fragment_link","input": "[foo](http://zulip.testserver/#narrow/channel/1-Denmark)",