Re: perl and utf8

[Anne Bennett <anne%porcupine.montreal.qc.ca@localhost>]

> I reported that Perl complained once I changed to a utf8 locale:

> > > My locale settings are:
(LANG, LC_CTYPE, LC_MESSAGES "en_CA.UTF-8", others "C")

> > > and "locale -a" *does* return en_CA.UTF-8 as one of its entries.

> I'm going to play with this a bit more and see if I can
> straighten it out a bit better.

Tests:

  fr_CA.ISO8859-15 for LANG, LC_CTYPE, LC_MESSAGES
    - no complaints from Perl
    - messages in French

  fr_CA.UTF-8 for LANG, LC_CTYPE, LC_MESSAGES
    - perl complains
    - messages are in English; oops
    - if I set LANG to 'C', perl stops complaining, but messages still
      in English

  fr_CA.UTF-8 for LC_CTYPE, fr_CA.ISO8859-1 for LC_MESSAGES, 'C' for rest
    - no complaints from Perl
    - messages in French but accented characters do not display
      correctly (not surprisingly)

My conclusions are:

  - The message catalogs are not yet available in UTF encoding for
    non-English languages (and for English, there are no non-ASCII
    characters anyway), so people who prefer another language are out of
    luck for now if they were hoping to switch to UTF8, at least at the
    O/S level -- I guess apps might still have their own, utf8-compliant
    message catalogs.

  - Perl doesn't like non-C "LANG", but has no problem with:
      LC_CTYPE="en_CA.UTF-8"
      LC_MESSAGES="en_CA.UTF-8"

Not that it matters much to me now whether I use en_CA or en_US,
at least until someone corrects the en_CA message catalog to read:

  No such file or directory, eh?!

;-)

Anne.