Bug#894217: [reprotest] kk_KZ.RK1048 again

[Ximin Luo]

Agustin Henze:
> [..]
> 
> Yeah, you're right about the test I sent before. I'm sorry I am not a Perl
> expert. But take a look at this, please
> 
> $ apt-get install locales-all -y # It's important having all locales installed
> because this is what reprotest does and also the way to reproduce the error
> 
> $ LC_ALL=kk_KZ.XXXXXX perl -e 'use open IO => ":locale"; print "yes\n"'
> perl: warning: Setting locale failed.
> perl: warning: Please check that your locale settings:
>         LANGUAGE = "de:de_DE",
>         LC_ALL = "kk_KZ.XXXXXX",
>         LANG = "de_DE.UTF-8"
>     are supported and installed on your system.
> perl: warning: Falling back to a fallback locale ("de_DE.UTF-8").
> yes
> 
> $ LC_ALL=kk_KZ.RK1048 perl -e 'use open IO => ":locale"; print "yes\n"'
> Cannot find encoding "RK1048" at /usr/share/perl/5.26/open.pm line 126.
> Cannot find encoding "RK1048" at /usr/share/perl/5.26/open.pm line 134.
> 
> $ LC_ALL=kk_KZ.RK1048 help2man
> Unknown encoding 'RK1048' at /usr/bin/help2man line 56.
> 

I'm sorry but I don't have any more time to look into this problem in depth, but this output suggests there is something wrong with either Perl or the locales-all package. Clearly it recognises RK1048 over XXXX but then something it's expecting is missing.

> [..]
> 
> Totally agree with you. But just to see if we are on the same page... You are
> saying that reprotest will fail with a lot of packages if you don't pass
> `--vary=-locales` because Perl and Python2 are broken, well I wouldn't say
> broken but they don't support the encoding that reprotest is using for testing
> random locales.
> 
> I thought that choosing another encoding in reprotest was reasonable because a
> lot of packages will fail if you try to use reprotest for this. We got 2 of 3
> packages broken by this.
> 

2 or 3 is not "a lot", back when I was doing reproducible builds full-time I would look at fixing issues affecting 20-30 packages or sometimes several hundred. Please try to understand the exact nature of the problem I am pointing out rather than coming back with the same point.

If you do more in-depth research on the problem I might be inclined to take it more seriously, but so far it looks like that isn't the case here. Reading the POSIX and C standards for locale-related functions might help you understand the problem better.

X

-- 
GPG: ed25519/56034877E1F87C35
GPG: rsa4096/1318EFAC5FBBDBCE
https://github.com/infinity0/pubkeys.git