Handle Parenthesis in links properly #358
Conversation
Links with parenthesis (like wikipedia links) are not correct when using autolinks Properly handle parenthesis in links. Fixes russross#291
| @@ -570,13 +570,16 @@ func TestInlineLink(t *testing.T) { | |||
|
|
|||
| // no closing ), autolinking detects the url next | |||
| "[disambiguation](http://en.wikipedia.org/wiki/Disambiguation_(disambiguation) is the", | |||
| "<p>[disambiguation](<a href=\"http://en.wikipedia.org/wiki/Disambiguation_(disambiguation\">http://en.wikipedia.org/wiki/Disambiguation_(disambiguation</a>) is the</p>\n", | |||
| "<p>[disambiguation](<a href=\"http://en.wikipedia.org/wiki/Disambiguation_(disambiguation)\">http://en.wikipedia.org/wiki/Disambiguation_(disambiguation)</a> is the</p>\n", | |||
There was a problem hiding this comment.
It changes the way unbalanced parenthesis is handled in links. This results in forming a correct link by giving priority to the parenthesis in the link. This however, does not conform(like earlier) to common mark processing http://spec.commonmark.org/dingus/?text=foo%20%5Bdisambiguation%5D(http%3A%2F%2Fen.wikipedia.org%2Fwiki%2FDisambiguation_(disambiguation)%20is%20the.
I was trying to make it follow the common mark spec but I could get my way around the code. 🤔
There was a problem hiding this comment.
Which section of Common Mark spec you had in mind here? It would be great to have this fix conform to Common Mark right away, let's try to see what it takes.
There was a problem hiding this comment.
See: http://spec.commonmark.org/0.27/#link-destination
A link destination consists of either
- a sequence of zero or more characters between an opening < and a closing > that contains no spaces, line breaks, or unescaped < or > characters, or
- a nonempty sequence of characters that does not include ASCII space or control characters, and includes parentheses only if (a) they are backslash-escaped or (b) they are part of a balanced pair of unescaped parentheses that is not itself inside a balanced pair of unescaped parentheses.
The spec does not talk about unescaped and unbalenced parenthesis in link destination. So I was just trying to make it behave the way http://spec.commonmark.org/dingus/ behaves for such links. It ignores the entire link processing in that case and renders it literally after processing. So:
foo [disambiguation](http://en.wikipedia.org/wiki/Disambiguation_(disambiguation) is the
becomes:
<p>foo [disambiguation](http://en.wikipedia.org/wiki/Disambiguation_(disambiguation) is the</p>.
There was a problem hiding this comment.
By not talking about unescaped and unbalanced parenthesis, I suppose, the commonmark spec regards those as invalid links and hence does not render them as HTML links.
There was a problem hiding this comment.
Looks like the section you're referring to talks about the "fully qualified" links, the ones with full link syntax.
And it seems that CommonMark only has autolinks with a hint: they require angle brackets around the link, which avoids the balancing problem altogether.
|
|
||
| if openDelim == 0 { | ||
| linkEnd-- | ||
| } |
There was a problem hiding this comment.
Won't this break this case?
foo (http://en.wikipedia.org/wiki/Disambiguation_(disambiguation)) bar
Autolink detector should recognize that the outer parentheses do not belong to the link. Can you add this case to the tests?
|
Since this pull is being ignored, here is a workaround: package main
import (
"bytes"
"github.com/yuin/goldmark"
"github.com/yuin/goldmark/extension"
)
func main() {
s1 := "https://en.wikipedia.org/wiki/The_Master_(2012_film)"
o1 := goldmark.New(goldmark.WithExtensions(extension.Linkify))
var b2 bytes.Buffer
o1.Convert([]byte(s1), &b2)
s2 := b2.String()
print(s2)
} |
ghost
mentioned this pull request
Links with parenthesis (like wikipedia links) are not correct when using autolinks
Properly handle parenthesis in links.
Fixes #291