From: Nguyen Thai Ngoc Duy <pclouds@gmail.com>
Date: Sat, 02 Jun 2012 10:55:55 GMT
Subject: Re: [PATCH 03/65] i18n: archive: mark parseopt strings for translation
Message-ID: <CACsJy8AGKQHKx9XU+baUagcmi6Sc-=faNshpED=Q46JpV_31WA@mail.gmail.com>
In-Reply-To: <20120529203759.GC22151@burratino>

On Wed, May 30, 2012 at 3:37 AM, Jonathan Nieder <jrnieder@gmail.com> wrote:
> Junio C Hamano wrote:
>> Nguyễn Thái Ngọc Duy  <pclouds@gmail.com> writes:
>
>>> --- a/archive.c
>>> +++ b/archive.c
>>> @@ -7,10 +7,10 @@
>>>  #include "unpack-trees.h"
>>>
>>>  static char const * const archive_usage[] = {
>>> -    "git archive [options] <tree-ish> [<path>...]",
>>> -    "git archive --list",
>>> -    "git archive --remote <repo> [--exec <cmd>] [options] <tree-ish> [<path>...]",
>>> -    "git archive --remote <repo> [--exec <cmd>] --list",
>>> +    N_("git archive [options] <tree-ish> [<path>...]"),
>>> +    N_("git archive --list"),
>>> +    N_("git archive --remote <repo> [--exec <cmd>] [options] <tree-ish> [<path>...]"),
>>> +    N_("git archive --remote <repo> [--exec <cmd>] --list"),
>>>      NULL
>>>  };
>>
>> I can understand the "help" text member of OPT_BLAH() construct, but
>> do we really want to translate the above?
>
> ...
>
> At least "git archive --list" should probably be kept untranslated
> to lessen the translator workload a bit.  Good catch.

Should it? On one hand, it's like 10 strings less for translator (I
fixed the series) but then we want to catch unmarked strings. _() will
always be called on "git archive --list" and will give fault positive
here. I intend to make "make test" warn every time it sees unmarked
strings, fault positives will be annoying.

Unless we have a notion to say "this string is deliberately kept
unstranslated", I think we better just mark them all.
-- 
Duy