bpo-34333: Fix %-formatting in Path.with_suffix()

[berkerpeksag]

https://bugs.python.org/issue34333

[ericvsmith]

LGTM

[berkerpeksag]

@ericvsmith thanks for the review! Do you have any suggestions on making the NEWS entry more descriptive? I'm pretty sure that I did a bad job explaining the bug :)

[ericvsmith]

I'm not sure it's much of an improvement, but had I been writing it from scratch I would have gone with:

Fix %-formatting in :meth:pathlib.PurePath.with_suffix when formatting an error message.

Although if I'd been doing it, I would have changed it to an f-string (of course!). I didn't suggest that just to minimize churn. But seriously: one of the motivating points of str.format (and later f-strings) was to get rid of exactly this kind of error. Maybe commit this fix (or not), then go through the entire file and use f-strings? There are a number of other error messages that seem to have this same problem, such as in both gethomedir methods. Not sure about others: the analysis is harder than just changing to f-strings. I realize this would make pre-3.6 backports harder, but that shouldn't still be an issue (barring finding a security-related problem).

Anyway, take that last suggestion or not, you won't hurt my feelings.

[berkerpeksag]

Thanks for the suggestions!

Initially, I thought about using f-strings in the whole module, but then I didn't want to introduce code churn and come up with artificial examples like the ones at https://bugs.python.org/issue27161 I'm planning to triage/review/merge pathlib issues and PRs in the coming weeks, so I can write proof-of-concept patch to implement your suggestion.

[miss-islington]

Thanks @berkerpeksag for the PR 🌮🎉.. I'm working now to backport this PR to: 3.7.
🐍🍒⛏🤖

[bedevere-bot]

GH-8731 is a backport of this pull request to the 3.7 branch.