MAINT: Ignore black format changes

[jarrodmillman]

See the comment below for an explanation of this: #713 (comment)

[choldgraf]

Can you please provide a little bit more context of the "what and the why"? I don't understand what this diff means, or why it is important to "ignore black format changes". It seems reasonable to me but I don't really know why I think that haha, I think it'd be helpful for others to understand and learn if we had a bit more info.

Is the goal here to prevent github from having noise about changes that were related to a bot? Is there some way we can automate this process? Or do we have to manually update it each time?

[jarrodmillman]

To see how this works, consider this commit

• 90550b66

That commit touched several files, but mostly applied autoformatting changes. When you use git blame at the commandline or if you use the blame view in the GitHub interface you get

See
https://github.com/pydata/pydata-sphinx-theme/blame/main/docs/conf.py

After merging this PR you would get

See
https://github.com/jarrodmillman/pydata-sphinx-theme/blame/git-blame-ignore-revs/docs/conf.py

In particular, you can see the difference by comparing lines 7--9.

I am not aware of any automated tool to do this, but it isn't a problem and the commits can be added to the file at any time.

I don't think this is super important. Personally, I don't have any trouble ignoring the handful of formatting commits when using git blame. But it is mildly helpful and may reduce concerns about using autoformatters like black, pyupgrade, autopep8, etc.

I would be happy if this was merged, but would also be happy if it wasn't.

[choldgraf]

I'm +1 on merging this, as long as we either:

• Can put a comment in the .git-blame-ignore-revs file that explains what it is for, maybe even just with a link to this PR.
• Put a little section in our contributing docs to document this.

I just don't want to add a dangling file that many people might not be familiar with without also adding a little bit of documentation :-)

[choldgraf]

Looks good to me, thanks for adding these extra explanations and links :-)