ed

Thanks for your reply :+1: Indeed, side margins on smaller viewports should be tweaked, but even more important is adjustable bottom margin for any width: it's a visually pleasant feature...

After doing some experimentation, it's a bad idea as it impairs the concurrency feature, which is the key. Waiting for other opinions (in case there is something to be done...

I have never worked on Windows so I won't be able to help you here. What you are looking for it the path to the hardware where MIDI signal flows...

Is it related to [this line](https://github.com/edouardtheron/py-midi/blob/master/midi/midi.py#L195)? ie, the object is falsly identifing a signal as beiing the end of a SysEx, although it is not?

Ok thanks for notifying. I won't have time to push a patch in the coming weeks, but a Pull Request is welcome :wink:

Thanks for your inquiry. I have another project that takes 100% of my productive time, so I won't be able to make any change to the current library. However, if...

Sorry for a late reply. If you try another type of MIDI message, does it work? What I want to know is: are you sure that the problem comes from...

I just tested with a smaller dataset, _ie_ instead of scanning all ~17k files, I only scan the first 50... And it works :thinking: Does it mean that the problem...

I managed to `scan/filter/sort/sink` the whole dataset by processing it by batches of 500 source files. ```python for i, batch in enumerate(batched(s3_urls, batch_size=500)): pl.scan_parquet( batch, ).filter( pl.col("date") < datetime.now() -...

I confirm that I still get the issue after upgrading to v1.0.0