gh-134887: Add references to locale module for locale-aware number formatting references in string module docs

[stefmolin]

Add references to locale module for locale-aware number formatting references in string module docs to avoid users thinking that the locale doesn't have a thousands separator for example.

• Issue: Documentation of n in the format-specifier should reference locale-setting #134887

[python-cla-bot]

All commit authors signed the Contributor License Agreement.

[ericvsmith]

I'm not sure we need to add a reference to setting the locale, since the text already says it's local aware.

[ericvsmith]

This isn't true, is it? I'm referring to the wording "must set": shouldn't it be "can set".

[stefmolin]

On my machine, nothing happens if I use n, without setting LC_NUMERIC, so it is "must set" for me. I'm happy to adjust the wording if there are cases where it is set, but I haven't seen it.

FWIW, I'm seeing that the C locale seems to be very minimal and doesn't have a thousands separator. Since that is what the default is, instead of the system locale, the thousands separator won't be set until the user sets LC_NUMERIC, meaning it is "must set" for everyone. I'm new to the C locale as well though, so let me know if that isn't the case.

[StanFromIreland]

Indeed, LC_NUMERIC is usually C, but I think using "must" is a little too absolute.

[willingc]

I think that is a fair assessment on "must". @stefmolin I think that breaking this into two sentences. The first that "Note that the system locale is not the default locale." And the second, "Depending on your use case, you may wish to set LC_NUMERIC."

[stefmolin]

I'm not sure we need to add a reference to setting the locale, since the text already says it's local aware.

While this may be obvious to a core CPython developer, it was not obvious to me or my friend, who is a Python trainer. We have each been using Python for many years. My friend routinely explores parts of the Python documentation to learn new things and share with the community. He found out about the n option and shared it on Bluesky, mentioning that the thousands separator may be nothing in your locale. I was excited to try this because I use the comma in the format specifier all the time and sometimes when I teach or show code snippets using this in places that use a different thousands separator, I get asked about this.

When I tried the n option, however, it didn't do anything on my machine, which I was shocked to see. I then read through the docs and couldn't see why it wouldn't work. The string module docs lack any reference to the locale module, which at this point I didn't even know existed. I searched on the Internet to see how I could check if my locale was correct, which led me to the locale module, and I was able to confirm it was set as I would expect. Since the locale was indeed set and the docs say that n provides locale-aware formatting, I was confused why it wasn't working.

I searched some more to discover why it wasn't using a thousands separator for the format. This time Stack Overflow led me to the real reason behind it not working: I needed to set the LC_NUMERIC category. I seriously doubt that my friend and I are the only people that don't find this obvious, so I made a PR in the hopes that the next person can figure out how to get this working while reading the docs directly, instead of having to troubleshoot it for a while (or just assuming it doesn't work).

[cmaureir]

I can reproduce this issue as well, and can confirm that even though my system has LC_NUMERIC the formatting with :n for f-strings doesn't properly work without setting the locale like: locale.setlocale(locale.LC_NUMERIC, 'en_US')

[github-actions]

This PR is stale because it has been open for 30 days with no activity.

[read-the-docs-community]

Documentation build overview

📚 cpython-previews | 🛠️ Build #32747484 | 📁 Comparing 5684a48 against main (85d3bcd)

  🔍 Preview build  

115 files changed · + 4 added · ± 111 modified

+ Added

• c-api/sentinel.html
• c-api/slots.html
• deprecations/pending-removal-in-3.21.html
• whatsnew/3.16.html

± Modified

• glossary.html
• index.html
• c-api/concrete.html
• c-api/dict.html
• c-api/exceptions.html
• c-api/extension-modules.html
• c-api/gcsupport.html
• c-api/index.html
• c-api/interp-lifecycle.html
• c-api/lifecycle.html
• and 101 more...

[willingc]

Thanks @stefmolin for the PR. I've restarted CI as we work through the PR together.

Hi @ericvsmith and @StanFromIreland, I'm mentoring Stefanie. She had a great chat with Guido at PyCon the other night. For a bit of background, she's given a bunch of tutorials at PyCons around the world and works in Bloomberg's security group.

Unless I hear any strong objections, I am going to approve and merge this PR since it is helpful.

[StanFromIreland]

I'm not sure why the CI run failed before, but it passed on a re-run.

[ericvsmith]

@willingc : no objections from me.

[miss-islington-app]

Thanks @stefmolin for the PR, and @encukou for merging it 🌮🎉.. I'm working now to backport this PR to: 3.13, 3.14, 3.15.
🐍🍒⛏🤖

[bedevere-app]

GH-150120 is a backport of this pull request to the 3.15 branch.

[bedevere-app]

GH-150121 is a backport of this pull request to the 3.14 branch.

[bedevere-app]

GH-150122 is a backport of this pull request to the 3.13 branch.