HTML comment handling doesn't match spec 0.31.2?

[robinst]

I'm updating commonmark-java to version 0.31.2 of the spec and was wondering about HTML comment handling. Namely, given this spec:

An HTML comment consists of <!-->, <!--->, or <!--, a string of
characters not including the string -->, and --> (see the
HTML spec).

I think these two should be parsed as comments:

foo <!-- test --->

foo <!-- test ---->

However, with the current dingus they're not (example 1, example 2):

<p>foo &lt;!-- test ---&gt;</p>

<p>foo &lt;!-- test ----&gt;</p>

The reason is probably this regex:

commonmark.js/lib/common.js

Line 27 in a4d859c

var HTMLCOMMENT = "<!-->|<!--->|<!--(?:[^-]|-[^-]|--[^>])*-->"

I think <!-->|<!--->|<!--.*?--> would work as expected.

[jgm]

OK, this looks like a holdover from our old XML-ish definition of comment.
Do you want to make PR?

[jgm]

I think cmark may also need to be fixed.
scanners.re has

  htmlcomment = "--" ([^\x00-]+ | "-" [^\x00-] | "--" [^\x00>])* "-->";

[robinst]

Do you want to make PR?

Yeah can do later. Do we want some additional examples in the spec as well?