Brainvision marker parser fails w/ [] characters

I have a (perhaps poorly chosen) marker in a file I received,

Mk9=Comment,Discontinuit? [B] 4156 ms,136544,1,0

which at brainvision.py:269 is being found with

items = re.findall(r"^Mk\d+=(.*)", mk_txt, re.MULTILINE)

but this regex stops completely at the [ character, not finding the rest of the line nor the markers which follow.

I’m not a regex ninja but if I manage to fix, I’ll send a PR. In the meantime, I will change remove the offending characters in the file and move on.

[sappelhoff]

You are right @maedoc: https://github.com/mne-tools/mne-python/blob/18755d22c3752b767dd493ffb4ffa5360bd322b4/mne/io/brainvision/brainvision.py#L305-L312

The r"\[.*\]" part is being used to find the line where a new section begins like [Marker Infos] or [Common Infos]

Actually, this points to a very simple fix. Instead of m = re.search(r"\[.*\]", mk_txt) we can use ^ and $ to get: m = re.search(r"^\[.*\]$", mk_txt). This asserts that [Marker Infos] and other section headers are always written isolated and on a single line (which is true in BrainVision format). Then even your weird markers should be captured 😉

see: #5412

[maedoc]

@sappelhoff indeed. I just realized that a few lines up from there, there’s a regex for r"\[.*\]" that may be the culprit.