NAME

Acme::Study::OS::DateLocales - study date-specific locales

SYNOPSIS

None. Just run the Pod.

DESCRIPTION

This module misuses the CPAN testers system to study the result of date-specific locale operations. Some of the questions to answer:

Can we rely on the fact that the locale implementation will return "bytes" in the given charset (encoding)?

Is the %OB extension of "strftime" in POSIX supported?

The results make me believe that one should not use POSIX-based locales for dates, but rather use DateTime::Locale.

RESULTS

* Solaris 10:
  * does not understand %OB
  * %B seems to return the genitive form
  * the "short" locale names seem to link to the non-utf8 forms (iso-8859-1 or so)
  * encoding seems to match the locale charset
  * all Serbian variants are cyrillic
* FreeBSD 6.2, 7.0, 7.2:
  * understands %OB, which is usually the nominative form of month names
  * %B has the genitive form (modulo bugs, see the Croatian locale)
  * encoding matches the locale charset
  * the ISO8859-2 variant of Serbian is latin, all others are cyrillic
  * it seems that all of locales are installed by default
* Linux (debian lenny):
  * does not understand %OB
  * encoding seems to match the locale charset
* Linux (debian etch):
  * does not understand %OB
  * %B returns the nominative form (at least for Croatian)
  * encoding seems to match the locale charset
  * the "short" locale names seem to link to the non-utf8 forms (iso-8859-1 or so)
* Linux (s390x-linux):
  * does not understand %OB
  * %B returns the nominative form (at least for Bosnian and Czech)
  * the @euro form seems to be the same like the "short" locale (that is, iso-8859-15 or so)
* OpenBSD 4.5:
  * does not seem to have the "locales -a" command, so only the default locale was tested
  * understands %OB, contents (genitive vs. nominative) unclear
* Darwin 8:
  * understands %OB, and seems to have the same bugs as the FreeBSD version
    (Croatian locale)
  * encoding matches the locale charset
  * the ISO8859-2 variant of Serbian is latin, all others are cyrillic
  * it seems that all of locales are installed by default
* MSWin32:
  * does not understand %OB
* cygwin:
  * does not seem to have the "locales -a" command, so only the default locale was tested
  * understands %OB, contents (genitive vs. nominative) unclear
* irix 6.5:
  * does not understand %OB
  * does not have utf-8 locales, but iso-8859-15 locales
  * the non-latin-1 locales (latin2, russian) don't have an encoding spec in its
    locale name, so detecting the encoding must be done through an heuristic

AUTHOR

Slaven Rezic.

SEE ALSO

Acme::Study::Perl.