brew.sh: fixes for UTF-8

[maxim-belkin]

• Have you followed the guidelines in our Contributing document?
• Have you checked to ensure there aren't other open Pull Requests for the same change?
• Have you added an explanation of what your changes do and why you'd like us to include them?
• Have you written new tests for your changes? Here's an example.
• Have you successfully run brew style with your changes locally?
• Have you successfully run brew tests with your changes locally?

---

Fix Homebrew detection mechanism of usable locale.

Order of locales #8047 (comment):

1. The user's current locale if it's UTF-8
2. C.UTF-8
3. en_US.UTF-8
4. First available UTF-8 locale
5. C

[MikeMcQuaid]

brew style is failing due to shellcheck: https://github.com/Homebrew/brew/pull/8072/checks?check_run_id=904644217

[maxim-belkin]

brew style is failing due to shellcheck: https://github.com/Homebrew/brew/pull/8072/checks?check_run_id=904644217

Yeh, my bad -- I noticed that but had to leave and left the PR without a note that I'll continue working on it.
By the way, brew style doesn't fail for me in a Homebrew/brew Docker container, which is weird.

[maxim-belkin]

Ready for a second review.

[MikeMcQuaid]

What's the action item for the user here?

[MikeMcQuaid]

(I'm wondering if this should be not output or be a hard-stop error).

[maxim-belkin]

install either a UTF-8 locale of choice or C locale.

[MikeMcQuaid]

In that case, yeh, I'd make this an odie with instructions on how to resolve.

[maxim-belkin]

sounds good to me but I'd like to run this by @sjackman 👀

[sjackman]

The user is pretty much guaranteed to have a C locale. If they do not, the world is burning. It's more likely that locale failed to run (they don't have this executable in their PATH, which itself is pretty unlikely) than they don't have a C locale. Rather than emit a warning or error in this else case, I'd just set LC_ALL=C and move on.

[maxim-belkin]

Rather than emit a warning or error in this else case, I'd just set LC_ALL=C and move on.

OK

[maxim-belkin]

Shall we remove c_regex then?

[sjackman]

Makes sense.

[sjackman]

Suggested change

	c_utf_regex='\bC\.(utf|UTF-)8'
	en_us_regex='en_US\.(utf|UTF-)8'
	utf_regex='[a-z][a-z]_[A-Z][A-Z]\.(utf|UTF-)8'
	c_regex='\bC\b'
	c_utf_regex='^C\.(utf8|UTF-8)$'
	en_us_regex='^en_US\.(utf8|UTF-8)$'
	utf_regex='^[a-z][a-z]_[A-Z][A-Z]\.(utf8|UTF-8)$'
	c_regex='^C$'

For style more than correctness. ^…$ is more common and easier to read than \b. Anchor all patterns for consistency. Super minor but I found (utf8|UTF-8) easier to read than (utf|UTF-)8.

[maxim-belkin]

^...$ don't work as one may expect because ^ and $ match the beginning and end of locale -a rather than the beginning and end of each word.

# In Homebrew/brew Docker image
$ locale -a
C
C.UTF-8
POSIX
en_US.utf8
$ locales=$(locale -a)
$ c_utf_regex='\bC\.(utf8|UTF-8)\b'
$ [[ $locales =~ $c_utf_regex ]] && echo ${BASH_REMATCH[0]}
C.UTF-8
$ c_utf_regex='^C\.(utf8|UTF-8)$'
$ [[ $locales =~ $c_utf_regex ]] && echo ${BASH_REMATCH[0]}
# nothing is printed; return status == 1

Anchor all patterns for consistency.

OK to use \b?

Super minor but I found (utf8|UTF-8) easier to read than (utf|UTF-)8.

OK.

[sjackman]

OK to use \b?

Yep! Thanks for the explanation.

[sjackman]

Thank you for this PR, Maxim!

[sjackman]

Thanks, Maxim!

[maxim-belkin]

Thanks for the reviews, @MikeMcQuaid and @sjackman!