Thank you, Akihiro-san.

I agree that it is better to use same variables as much as possible.

Akihiro Motoki wrote on 9/19/2016 7:21 AM:
Hi Ian,

gettext itself allows this pattern [1] but django compilemessage
expects variables with same names are used in all plural forms (See
Note in [2]).
IMHO it is better to use same variables in all plural forms in code as
possible as we can.
As translator, I try to use the second pattern (i.e., plural form) as
source string as a plural form contains sufficient and full pattern in
most cases.
Yep it would be a Fuzzy form for translation in Zanata, but that's okay for me. I have fixed in Korean,
and checked that there is no error for 'msgfmt --check-format ' command.


With many thanks,

/Ian
What I am not sure is this happens only for languages with nplural==1
(like Korean, Japanese and so on).

Akihiro

[1] 
https://www.gnu.org/software/gettext/manual/html_node/Translating-plural-forms.html#Translating-plural-forms
[2] 
https://docs.djangoproject.com/en/1.10/topics/i18n/translation/#pluralization


2016-09-16 18:57 GMT+09:00 Ian Y. Choi <ianyrc...@gmail.com>:
Hello,

Recently, thanks to the help from Andreas, I am investigating on a broken
job
for the translation import of Horizon project [1].
The actual error message is
: openstack_dashboard/locale/ko_KR/LC_MESSAGES/django.po:8273: a format
specification for argument 'req' doesn't exist in 'msgstr[0]'

And then I have found that this error was from a string in horizon -
openstack_dashboard [2]

             error_message = ungettext_lazy(
                 'The requested instance cannot be launched as you only '
                 'have %(avail)i of your quota available. ',
                 'The requested %(req)i instances cannot be launched as you '
                 'only have %(avail)i of your quota available.',
                 count)
             params = {'req': count,
                       'avail': available_count}

In i18n translation infrastructure, only one of two (for singular and
plural) strings should be
selected, translated, saved, and finally imported back to Horizon project as
[1].
However, the first string only used "%(avail)i" string variable,
and the second string used both "%(req)i" and "%(avail)i" string variables.

Since one Korean translator selected the first string, there will be no
"%(req)i" in Korean po file, which generates such an error.
So my suggestion is to use the same string variables for both two strings
when we use ungettext_lazy function.

Is it a bug from Horizon? Would it be other issues when we change like
: from 'The requested instance cannot be launched as you only ' to 'The
requested %(req)i instance cannot be launched as you only '?


[1]
http://logs.openstack.org/periodic/horizon-propose-translation-update/c038aab/console.html#_2016-09-16_07_37_22_312092
[2]
http://git.openstack.org/cgit/openstack/horizon/tree/openstack_dashboard/dashboards/project/instances/workflows/create_instance.py#n214


With many thanks,

/Ian

_______________________________________________
OpenStack-I18n mailing list
openstack-i...@lists.openstack.org
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-i18n



__________________________________________________________________________
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

Reply via email to