Re: i18n, part 2

["Eli Zaretskii" <eliz@gnu.org>]

> Date: Wed, 15 Dec 2004 01:07:45 +0100
> From: Baurjan Ismagulov <ibr@ata.cs.hun.edu.tr>
> 
> here's the new version. I've addressed two issues:
> 
> * Providing context.
> 
>   I've followed the advice from Andreas Schwab and provided examples of
>   the whole phrases as they are printed by gdb. In this way, the code is
>   readable, and the message file has even more context than is provided
>   by the translated string itself. I've done this only for permutations;
>   obvious cases like "big endian" / "little endian" have been rewritten.
>   The relevant excerpt from the output of "xgettext -o a.po -k_
>   --add-comments=i18n *.c" is attached.

Thanks.  This version is okay with me, provided that we record
somewhere that when the flag day comes, xgettext will be invoked
henceforth with the --add-comments=i18n switch.  I'm unsure where to
put that info.  Andrew?

>    if (arm_fp_model == ARM_FLOAT_AUTO 
>        && gdbarch_bfd_arch_info (current_gdbarch)->arch == bfd_arch_arm)
> -    printf_filtered ("  - the default for the current ABI is \"%s\".\n",
> +    printf_filtered (_("  - the default for the current ABI is \"%s\".\n"),
>  		     fp_model_strings[tdep->fp_model]);

This lacks the leading context, which makes it unclear what default is
this string about.  Can we either rewrite the printf to include the
context, or resolve it with a i18n comment?