[publican-list] Use of language-codes in Publican

Asgeir Frimannsson asgeirf at redhat.com
Tue Sep 30 04:57:59 UTC 2008


----- "Asgeir Frimannsson" <asgeirf at redhat.com> wrote:
> Hi publicans and others,
> 
> I was doing some work earlier today with Fedora release notes and
> publican, and noticed a potential issue with locale names. 
> 
> Publican seem to expect locales in the form of $lang-$country, e.g.
> en-US or hi-IN. 
> 
> 1) Many open source projects use underscore rather than hyphen for
> locale codes, for example en_US or hi_IN. I'm all in favour of the
> publican-approach, and converting projects to use these is trivial.
> 
> 2) For some languages, a country code in the locale name is not used,
> and this breaks publican when running 'make report-all'. 

Sorry, the target that fails is 'publish-report', where the process spits out a lot of messages like:
"Use of uninitialized value $language in string eq at /usr/bin/StSe_Reports line 72."

cheers,
asgeir

>Not sure if
> other parts of publican breaks. I'm not convinced that enforcing a
> country-code in the locale name is the right thing to do. Should this
> be considered a bug? The country-code (additional sub-tags) in RFC3066
> [1] is optional.
> 
> There is also a special case for e.g. Serbian (sr and sr_Latn). I
> think using RFC3066 these would be written sr and sr-Latn, or with a
> country code sr-RS and sr-Latn-RS.
> 
> cheers,
> asgeir
> 
> [1] http://www.ietf.org/rfc/rfc3066.txt
> 
> _______________________________________________
> publican-list mailing list
> publican-list at redhat.com
> https://www.redhat.com/mailman/listinfo/publican-list
> Wiki: https://fedorahosted.org/publican




More information about the publican-list mailing list