Possibly faulty parsing of close references

[garazdawi]

The parser seems to think that [a] [b] is the same as [a][b].

iex(1)> EarmarkParser.as_ast "[a] [b]\n[a]: a\n[b]: b"
{:ok, [{"p", [], [{"a", [{"href", "b"}, {"title", ""}], ["a"], %{}}], %{}}], []}

The standard is a bit undecided. Gruber Markdown seems to do the same interpretation as earmark_parser, while github (and in extentions commonmark) thinks that it should be treated as two separate references.

I'm opening this issue to just track the discepency and see if you think we should do something about it.

[RobertDober]

Ty Lukas
well I should take this into account when (or rather if) I will ever come to write version 1.5 which aims for more GFM or 2.0 which should be compliant, right now things are looking not good.

But maybe this is a quick fix? I'll keep you posted

[garazdawi]

I noticed it while reading our release notes where we list which issues and PRs a certain change was introduced by. It was an easy fix on our side and the scripts that generate the information have been updated to always insert a comma, so no rush on our end.

[RobertDober]

Good because while GFM compliance is more or less expected it is always risky to make patch releases which change behviour (on extreme edge case, but still).

Can I close it than? leaving open for further reference

[RobertDober]

My bad no, I'll keep it open