From: Robert Griesemer <gri@golang.org>
Date: Tue, 8 Apr 2014 20:51:44 +0000 (-0700)
Subject: go/doc: fix URL matched in ToHTML
X-Git-Tag: go1.3beta1~161
X-Git-Url: http://www.git.cypherpunks.su/?a=commitdiff_plain;h=9610b616c6936edde5aa54312c8dc5e182aebf73;p=gostls13.git

go/doc: fix URL matched in ToHTML

Permit paired parentheses in URLs such as:

http://en.wikipedia.org/wiki/Camellia_(cipher)

Fixes #5043.

LGTM=rsc
R=rsc
CC=golang-codereviews
https://golang.org/cl/85610043
---

diff --git a/src/pkg/go/doc/comment.go b/src/pkg/go/doc/comment.go
index 274a625cf0..f414ca4090 100644
--- a/src/pkg/go/doc/comment.go
+++ b/src/pkg/go/doc/comment.go
@@ -45,13 +45,13 @@ func commentEscape(w io.Writer, text string, nice bool) {
 
 const (
 	// Regexp for Go identifiers
-	identRx = `[a-zA-Z_][a-zA-Z_0-9]*` // TODO(gri) ASCII only for now - fix this
+	identRx = `[\pL_][\pL_0-9]*`
 
 	// Regexp for URLs
-	protocol = `(https?|ftp|file|gopher|mailto|news|nntp|telnet|wais|prospero):`
+	protocol = `https?|ftp|file|gopher|mailto|news|nntp|telnet|wais|prospero`
 	hostPart = `[a-zA-Z0-9_@\-]+`
-	filePart = `[a-zA-Z0-9_?%#~&/\-+=]+`
-	urlRx    = protocol + `//` + // http://
+	filePart = `[a-zA-Z0-9_?%#~&/\-+=()]+` // parentheses may not be matching; see pairedParensPrefixLen
+	urlRx    = `(` + protocol + `)://` +   // http://
 		hostPart + `([.:]` + hostPart + `)*/?` + // //www.google.com:8080/
 		filePart + `([:.,]` + filePart + `)*`
 )
@@ -73,6 +73,29 @@ var (
 	html_endh   = []byte("</h3>\n")
 )
 
+// pairedParensPrefixLen returns the length of the longest prefix of s containing paired parentheses.
+func pairedParensPrefixLen(s string) int {
+	parens := 0
+	l := len(s)
+	for i, ch := range s {
+		switch ch {
+		case '(':
+			if parens == 0 {
+				l = i
+			}
+			parens++
+		case ')':
+			parens--
+			if parens == 0 {
+				l = len(s)
+			} else if parens < 0 {
+				return i
+			}
+		}
+	}
+	return l
+}
+
 // Emphasize and escape a line of text for HTML. URLs are converted into links;
 // if the URL also appears in the words map, the link is taken from the map (if
 // the corresponding map value is the empty string, the URL is not converted
@@ -92,18 +115,26 @@ func emphasize(w io.Writer, line string, words map[string]string, nice bool) {
 		// write text before match
 		commentEscape(w, line[0:m[0]], nice)
 
-		// analyze match
+		// adjust match if necessary
 		match := line[m[0]:m[1]]
+		if n := pairedParensPrefixLen(match); n < len(match) {
+			// match contains unpaired parentheses (rare);
+			// redo matching with shortened line for correct indices
+			m = matchRx.FindStringSubmatchIndex(line[:m[0]+n])
+			match = match[:n]
+		}
+
+		// analyze match
 		url := ""
 		italics := false
 		if words != nil {
-			url, italics = words[string(match)]
+			url, italics = words[match]
 		}
 		if m[2] >= 0 {
 			// match against first parenthesized sub-regexp; must be match against urlRx
 			if !italics {
 				// no alternative URL in words list, use match instead
-				url = string(match)
+				url = match
 			}
 			italics = false // don't italicize URLs
 		}
diff --git a/src/pkg/go/doc/comment_test.go b/src/pkg/go/doc/comment_test.go
index 9f29a61153..ad65c2a27f 100644
--- a/src/pkg/go/doc/comment_test.go
+++ b/src/pkg/go/doc/comment_test.go
@@ -148,13 +148,16 @@ func TestToText(t *testing.T) {
 }
 
 var emphasizeTests = []struct {
-	in  string
-	out string
+	in, out string
 }{
 	{"http://www.google.com/", `<a href="http://www.google.com/">http://www.google.com/</a>`},
 	{"https://www.google.com/", `<a href="https://www.google.com/">https://www.google.com/</a>`},
 	{"http://www.google.com/path.", `<a href="http://www.google.com/path">http://www.google.com/path</a>.`},
+	{"http://en.wikipedia.org/wiki/Camellia_(cipher)", `<a href="http://en.wikipedia.org/wiki/Camellia_(cipher)">http://en.wikipedia.org/wiki/Camellia_(cipher)</a>`},
 	{"(http://www.google.com/)", `(<a href="http://www.google.com/">http://www.google.com/</a>)`},
+	{"http://gmail.com)", `<a href="http://gmail.com">http://gmail.com</a>)`},
+	{"((http://gmail.com))", `((<a href="http://gmail.com">http://gmail.com</a>))`},
+	{"http://gmail.com ((http://gmail.com)) ()", `<a href="http://gmail.com">http://gmail.com</a> ((<a href="http://gmail.com">http://gmail.com</a>)) ()`},
 	{"Foo bar http://example.com/ quux!", `Foo bar <a href="http://example.com/">http://example.com/</a> quux!`},
 	{"Hello http://example.com/%2f/ /world.", `Hello <a href="http://example.com/%2f/">http://example.com/%2f/</a> /world.`},
 	{"Lorem http: ipsum //host/path", "Lorem http: ipsum //host/path"},
@@ -171,3 +174,34 @@ func TestEmphasize(t *testing.T) {
 		}
 	}
 }
+
+var pairedParensPrefixLenTests = []struct {
+	in, out string
+}{
+	{"", ""},
+	{"foo", "foo"},
+	{"()", "()"},
+	{"foo()", "foo()"},
+	{"foo()()()", "foo()()()"},
+	{"foo()((()()))", "foo()((()()))"},
+	{"foo()((()()))bar", "foo()((()()))bar"},
+	{"foo)", "foo"},
+	{"foo))", "foo"},
+	{"foo)))))", "foo"},
+	{"(foo", ""},
+	{"((foo", ""},
+	{"(((((foo", ""},
+	{"(foo)", "(foo)"},
+	{"((((foo))))", "((((foo))))"},
+	{"foo()())", "foo()()"},
+	{"foo((()())", "foo"},
+	{"foo((()())) (() foo ", "foo((()())) "},
+}
+
+func TestPairedParensPrefixLen(t *testing.T) {
+	for i, tt := range pairedParensPrefixLenTests {
+		if out := tt.in[:pairedParensPrefixLen(tt.in)]; out != tt.out {
+			t.Errorf("#%d: mismatch\nhave: %q\nwant: %q", i, out, tt.out)
+		}
+	}
+}