EEGLab bug writing HDF5 persisting in MNE-BIDS

[christinerogers]

Description of the problem

Opening this issue as a follow-on from #652 which was resolved/closed with a workaround recommendation #653 specific to Matlab .SET files -- as discussed with @sappelhoff in the BIDS Maintainers meeting July 13, 2023

Given .Set is one of only 4 supported file types for BIDS EEG and is extremely widespread (with EEGLab use) in the community --
Although the workaround is effective, my MCIN group (makers of EEG2BIDS converter and EEGNet.org at McGill University would ideally like to see a solution factored into MNE-BIDS if possible

Thanks, and let us know if details / examples would help, or how we can get involved.

cc @larsoner for EEGNet Technical Committee and MCIN developers @jeffersoncasimir @laemtl

Steps to reproduce

please see #652

Expected results

please see #652

Actual results

please see #652

Additional information

please see #652

[larsoner]

Based on a very quick reading it seems like there was a fix put in #653 that has been merged and released. Is there an updated minimal example we can work from to fix the issue?

[sappelhoff]

@christinerogers we would indeed need a detailed report of the issues you are facing, including MWEs, please

[sappelhoff]

closing, as no reproducible error has been reported -- feel free to reply with instructions on how to reproduce an error / a problem, so that we can fix it.

In the meantime, I assume that #1187 fixes some problems.