Markdown links containing [ and ] are broken

[RagnarGrootKoerkamp]

The library doesn't allow [ and ] inside a markdown link text to occur, while the spec allows:

• matching, unescaped [ and ]
• escaped \[ and \].

(I came here via https://gitlab.com/gitlab-org/gitlab-ce/issues/66601, where the formatting of commit messages is broken when the message includes [ and ]. I'd try to solve it on the gitlab side, but that's only possible when escaped brackets are picked up correctly.)

[stevebissett]

@RagnarGrootKoerkamp - Did you come up with a solution for this. I am also having this issue at the moment.

[RagnarGrootKoerkamp]

https://gitlab.com/gitlab-org/gitlab/-/issues/31263#note_215961096
That fixed it for me at the time: I just create slack formatted links directly in gitlab instead of having markdown links converted to slack links at a later point.

Not sure if it still works though; I lost those local changes after updating and didn't bother to make them again.

[RagnarGrootKoerkamp]

Here's a page explaining a general regex for matching links in markdown: https://medium.com/@michael_perrin/match-markdown-links-with-advanced-regex-features-fc5f9f4122bc

It uses the following for matching the [..] part:

(?<text_group>                  # Text group, including square brackets
  \[
    (?>                         # (?> defines an atomic group, this is a performance improvement when using recursion
      [^\[\]]+                  # Look for any char except closing square bracket
      |(?&text_group)           # OR: find recursively an other pattern with opening and closing square brackets
    )*
  \]
)

I'd hope that putting that bit of code here should do the trick.

I'd do it myself but don't have any Ruby experience. Maybe you can test it and create a PR @stevebissett

[stevebissett]

Thanks @RagnarGrootKoerkamp - I landed up working around this by going the conversion before using the Slack formatter, because it was easy to do so in my use case.

I did have a go at trying to adjust the regex's and markdown handling to support this but I haven't managed to find working code that makes the spec pass. I am not all that familiar with Regex's at this complexity.

I have pushed up a spec for it here:
#120

I did have a look at some other Ruby markdown converters like redcarpet and the RDoc markdown to try and get a better idea of some alternative approaches but didn't come right.