Danek Duvall-san wrote (10/29/08 11:10 AM):
> On Wed, Oct 29, 2008 at 10:40:55AM +0900, Takao Fujiwara - Tokyo S/W Center 
> wrote:
> 
>> Danek Duvall-san wrote (10/29/08 03:11 AM):
>>> I don't understand why we need to call bindtextdomain() at all, but that
>>> functionality should be subsumed by the GNUTranslations object.  Calling
>>> textdomain() shouldn't be necessary for the same reason.  And gettext()
>>> doesn't seem to provide any benefit.
>> As I explained, gettext.textdomain(None) is needed in several situations 
>> and also bindtextdomain(foo, None).
>> The default gettext.install doesn't seems to provide this. So I think using 
>> textdomain is better.
> 
> This doesn't explain anything to me.  You don't tell me why it's needed,
> and why I'm wrong in my analysis.  Nor do you explain why the code change
> you propose is inconsistent in its application across all modules.

OK, I see. I thought you were convinced since you didn't reply the previous 
mail.

When the argument of C-version textdomain(3) is NULL, the behavior is defined 
and then the compatible functions are needed in Python applications.

E.g.
current_domain = textdomain (None)
open (something + current_domain + ".txt", "r")
print dgettext (current_domain + "-lib", "msgid")


> 
> "Doesn't seem" and "I think" don't convince me that you're right, just that
> you're not really sure what's going on and you happened to stumble on
> something that appeared to work.

OK, my expression might be wrong.
gettext.install doesn't touch _current_domain in gettext.py so 
gettext.textdomain(None) doesn't return the current domain.

Does it make sense?

Thanks,
fujiwara

> 
> If that's not the case, great.  But please deconstruct my analysis.
> 
> Thanks,
> Danek
> 

_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss

Reply via email to