don't allow | as part of a URL

This was an unintended bug added ages ago. In c64477d, we modified
midChar to add \p{Po} to the character set. Unfortunately, we added |,
essentially ending up with [somechars|\p{Po}]. In other words, | was now
part of the set, while we only meant it as a regex "or" operator.

Fix this, and add a test.

Fixes #26.

Author: mvdan

xurls.go

@@ -21,7 +21,7 @@ const (
 	otherSymb = `\p{So}`
 	endChar   = iriChar + `/\-+_&~*%=#` + currency + otherSymb
 	otherPunc = `\p{Po}`
-	midChar   = endChar + `|` + otherPunc
+	midChar   = endChar + otherPunc
 	wellParen = `\([` + midChar + `]*(\([` + midChar + `]*\)[` + midChar + `]*)*\)`
 	wellBrack = `\[[` + midChar + `]*(\[[` + midChar + `]*\][` + midChar + `]*)*\]`
 	wellBrace = `\{[` + midChar + `]*(\{[` + midChar + `]*\}[` + midChar + `]*)*\}`

xurls_test.go

@@ -104,7 +104,7 @@ var constantTestCases = []testCase{
 	{`'http://foo.com/bar'`, `http://foo.com/bar`},
 	{`'http://foo.com/bar'more`, `http://foo.com/bar'more`},
 	{`"http://foo.com/bar"`, `http://foo.com/bar`},
-	{`http://a.b/a0/-+_&~*%=#@.,:;'?!|[]()a`, true},
+	{`http://a.b/a0/-+_&~*%=#@.,:;'?![]()a`, true},
 	{`http://a.b/a0/$€¥`, true},
 	{`http://✪foo.bar/pa✪th©more`, true},
 	{`http://foo.bar/path/`, true},
@@ -126,6 +126,7 @@ var constantTestCases = []testCase{
 	{`http://foo.bar/path!`, `http://foo.bar/path`},
 	{`http://foo.bar/path@`, `http://foo.bar/path`},
 	{`http://foo.bar/path|`, `http://foo.bar/path`},
+	{`http://foo.bar/path|more`, `http://foo.bar/path`},
 	{`http://foo.bar/path<`, `http://foo.bar/path`},
 	{`http://foo.bar/path<more`, `http://foo.bar/path`},
 	{`http://foo.com/path_(more)`, true},