Oh, good point, I didn't compare it to the other code blocks in that file. Yeah, I suppose it would be slightly more consistent, although I see two, three, four, and five spaces of indentation relative to the preceding paragraphs... It seems like three is the most common though, and that's what the dev guide says as well (three spaces in reST, 4 spaces inside the Python blocks, but the indentation of the code blocks would be part of reST syntax). I'll normalise the entire document to use three for consistency.

I don’t think the churn is worth it. This will not change the output, nor help developers.

I mostly agree with @merwok -- the churn is too large, and in most cases it does not affect the output. But in two cases, in the original code example and in the versionchanged directive for asynccontextmanager, incorrect indentation affects the output. It is a bug which should be fixed. I may miss other such cases, please re-check me.

@JustAnotherArchivist Are you still interested in combing through the docs for whitespace / visible formatting changes? I think the important ones listed here are the original stray space and also some three-space indentation in code examples. Another possible change is to remove the >>> from code examples as it doesn't help much with copying:

Another possible change is to remove the >>> from code examples as it doesn't help much with copying:

I think that’s been addressed quite some time ago: there is now a formatting that adds buttons to toggle prompts on/off, and easily copy the code block.

I think that’s been addressed quite some time ago: there is now a formatting that adds buttons to toggle prompts on/off, and easily copy the code block.

Not 100% on how it works, but in the case above it doesn't have this button in the top right:

I'm guessing that all of the code example would need the >>>/... for the button to be generated, otherwise it thinks it's a regular code block.

I'll start on a new PR here then and close this when that one's open - let me know if you'd still like to continue this one.

See #98111

Apologies, I hadn't seen the notification emails for this for some reason. I disagree with the 'churn' argument, but that's a bit late now. Thanks for the fix, @slateny!

No worries, thanks for the update 🙂