[pacman-dev] weird error when makepkg

[甘露(Gan Lu)]

On Thu, Jun 24, 2010 at 8:16 PM, Dan McGee <dpmcgee at gmail.com> wrote:
> 2010/6/24 甘露(Gan Lu) <rhythm.gan at gmail.com>:
>> 2010/6/24 Allan McRae <allan at archlinux.org>:
>>> On 24/06/10 18:02, 甘露(Gan Lu) wrote:
>>>> there are two reports related to my translation (Chinese simplified),
>>>> but I can't find any similar items in my po file, so I ask here for
>>>> help:
>>>> 1. when makepkg with -sifr， there are error in finishing:
>>>> ==>  完成创建：plymouth-git 20100624-1 (2010年 06月 24日 星期四 15:48:46 CST)
>>>> /usr/bin/makepkg: line 93: printf: `$': invalid format character
>>>> the correspondent line in /usr/bin/makepkg is:  printf
>>>> "${GREEN}==>${ALL_OFF}${BOLD} ${mesg}${ALL_OFF}\n" "$@">&2
>>>> sorry for Chinese but this error only occurs in zh_CN environment, C
>>>> or en_US.UTF-8 works well.
>>
>> When with -c argue, makepkg works well and message is normal like:
>>
>> ==> 完成创建：plymouth-git 20100624-1 (2010年 06月 24日 星期四 20:03:06 CST)
>> ==> 正在清理...
>>
>> Could this be possibly related to gettext not support %1$s format anymore?
>
> You've made this very same mistake before, command line gettext
> doesn't support position parameters...
> http://projects.archlinux.org/pacman.git/commit/?id=5ffc1ad3cdb1ec252739c62c9d614aba468bb22c
Oh, I know that, in fact I don't change all translation but only that
one, there are eventually many such translation existing in Chinese po
file, I will try to fix all later, however I am still wonder why -c
argue doesn't be affected?
>
>>> What is the line normally output after this?  That is the line that is
>>> failing.
>>>
>>>> 2. other one is (report by others, I currently don't know what argue
>>>> he use 'makepkg'):
>>>> /usr/bin/makepkg: line 902: * 1024 : syntax error: operand expected
>>>> (error token is "* 1024 ")
>>>> the correspondent line in /usr/bin/makepkg is:  size="$((
>>>> ${size%%[^0-9]*} * 1024 ))"
>>>
>>> This is due to the cw package.  That provides a "du" with coloured
>>> output that breaks the calculation of the size.
>> Is this a bug?
>
> http://bugs.archlinux.org/task/19865
> http://bugs.archlinux.org/task/19932
>
>