Fix <em> issue with mixed content #1410

[x13machine]

Marked version:

Markdown flavor: Markdown.pl|CommonMark|GitHub Flavored Markdown|n/a

Description

• Fixes this bug Incompatibility between links wrapped in asterisks, pedantic, and GFM #1410
• I added a test case for this bug.

It was caused _ and * parsed being differently in the em regex. I just replaced the part that searches for *test* and replaced it with the part searches for _test_ and then just charged the characters. I don't know why they were different. Maybe there's a reason for it. All the tests pass, so I dunno.

Contributor

• Test(s) exist to ensure functionality and minimize regression

Committer

In most cases, this should be a different person than the contributor.

• Draft GitHub release notes have been updated.
• CI is green (no forced merge required).
• Merge PR

[x13machine]

ran wrong test command

[x13machine]

Wait, is it expecting a bug to be there? If it is, I seemed to have fixed it.

[UziTech]

Ya the spec tests that are currently failing will fail if they start passing. You will need to remove shouldFail: true from /test/specs/commonmark/commonmark.0.28.json line #3730

It looks like you are also failing some tests that used to pass.

[UziTech]

The shouldFail lets us know which tests are fixed by each PR and let us keep track of the spec tests that we need to work on.

[x13machine]

ok I figured out why they are different. It was to prevent it from being interpreted as an ordered List.

Also the underscore group has a bug: https://marked.js.org/demo/?text=_%3Cimg%20src%3D%22https%3A%2F%2Fwww.google.com%2Ffavicon.ico%22%20title%3D%22*_%22%2F%3E%0A%0A*%3Cimg%20src%3D%22https%3A%2F%2Fwww.google.com%2Ffavicon.ico%22%20title%3D%22*%22%2F%3E&options=&version=master

The asterisk was processed correctly. While the underscore was not. So the bug was transferred over.

So the underscore and asterisk groups need to be rewritten.

[x13machine]

Ok I think, I fixed it this time.

[UziTech]

@davisjam could you check if these regexps are secure from redos?

Here are the new and old derived regexps for em:

New:

/^_([^\s_])_(?!_)|^\*([^\s*<\[])\*(?!\*)|^_([^\s<][\s\S]*?[^\s_])_(?!_|[^\s!"#$%&'()*+,\-.\/:;<=>?@\[^_{|}~])|^_([^\s_<][\s\S]*?[^\s])_(?!_|[^\s!"#$%&'()*+,\-.\/:;<=>?@\[^_{|}~])|^\*([^\s<"][\s\S]*?[^\s\*])\*(?!\*|[^\s!"#$%&'()*+,\-.\/:;<=>?@\[^_{|}~])|^\*([^\s*"<\[][\s\S]*?[^\s])\*(?!\*)/

Old:

/^_([^\s_])_(?!_)|^\*([^\s*"<\[])\*(?!\*)|^_([^\s][\s\S]*?[^\s_])_(?!_|[^\s!"#$%&'()*+,\-.\/:;<=>?@\[^_{|}~])|^_([^\s_][\s\S]*?[^\s])_(?!_|[^\s!"#$%&'()*+,\-.\/:;<=>?@\[^_{|}~])|^\*([^\s"<\[][\s\S]*?[^\s*])\*(?!\*)|^\*([^\s*"<\[][\s\S]*?[^\s])\*(?!\*)/

[UziTech]

@x13machine could you rebase this so we could get it merged?

[x13machine]

Ok I'll do that today

[x13machine]

How do I fix the errors? 😕

[UziTech]

looks like example 455 should be "shouldFail": true and the "shouldFail" line should be removed on example 418 and 477

Also move the new test files to /test/specs/new/

[UziTech]

I'm not sure why the latest commit didn't trigger travis to run the tests, but I ran it locally and everything passed.

[UziTech]

This still doesn't fix *[link*](url* but that can be a separate PR.

marked demo
commonmark demo