Fix tag RegEx

[MicaelJarniac]

Type of changes

• Bug fix
• New feature
• Documentation / docstrings
• Tests
• Other

Checklist

• I've run the latest black with default args on new code.
• I've updated CHANGELOG.md and CONTRIBUTORS.md where appropriate.
• I've added tests for new code.
• I accept that @willmcgugan may be pedantic in the code review.

Description

Closes #3024

Updated regex pattern to improve color formatting. The tag_contents group was updated to exclude < character ([^<]), preventing the matching of tags that contain other tags inside. Additionally, a *? was used for a non-greedy match, making the group match as little as possible, thus avoiding overreaching matches.

[MicaelJarniac]

I'll be trying to write tests that check if this fix is working.

[MicaelJarniac]

Nevermind. I could not find where this kind of test is.

[MicaelJarniac]

Before

Non-greedy

Final

[MicaelJarniac]

Apparently, it's still buggy, only less so.

[MicaelJarniac]

from rich.pretty import pprint

class A:
    def __init__(self, a=None):
        self.a = a

    def __repr__(self) -> str:
        return f"A(a={self.a!r}, b=1.23, c=None, d=(1, 2))"

pprint([A("<"), A(">")])

[willmcgugan]

You're not always going to be able to parse things perfectly. REPRs aren't a formal grammar, and there will always be output which break it.

Happy to accept a PR that improves on it though. One thing to be aware of is performance. You might print a lot of text with Rich, and overly expensive regexes can be a problem.

[MicaelJarniac]

I'll see if there's an easy fix for this second problem I've encountered. The fix for the first one seems to work alright and is already on this PR.

[MicaelJarniac]

Well, simply removing the tag_contents group entirely does fix it. I'll see what other side effects it has. The style for repr.tag_contents is just Style(color="default").

[MicaelJarniac]

I found where the tests for this are, and will try to adapt them.
I'll make this PR a draft for now, the tests need some work and there's this other problem I just discovered today.

[MicaelJarniac]

Found a bug with my implementation regarding recursive tags. This should do the trick, but I'm yet to actually test it:

(?P<tag_start><)(?P<tag_name>[-\w.:|]*)(?P<tag_contents>[^><]+|(?R))*+(?P<tag_end>>)

[MicaelJarniac]

Before

After

[MicaelJarniac]

Nevermind, it seems like Python's re doesn't support recursion.

[MicaelJarniac]

What I'm currently doing to patch it locally is this:

# Fix Rich repr tags
# https://github.com/Textualize/rich/pull/3025
from rich.default_styles import DEFAULT_STYLES
from rich.highlighter import ReprHighlighter
from rich.style import Style

DEFAULT_STYLES["repr.tag_contents"] = Style.null()

ReprHighlighter.highlights[0] = r"(?P<tag_start><)(?P<tag_name>[-\w.:|]*)(?P<tag_contents>[^<]*?)(?P<tag_end>>)"

[willmcgugan]

Closing, assumed stale. But if you are still working on it, feel free to reopen.