International APPs / pygettext

[Martin v. Löwis]

Ingo Linkweiler <i.linkweiler at web.de> writes:

> msgfmt: messages.de.po: warning: Charset "CHARSET" is not a portable
> encoding name. Message conversion to user's charset might not work.

I recommend to use iso-8859-1 here.

> I ignored this warning and got a file "messages.mo"
> 
> Now I run the python source. Result :-(
> 
> Hello World

You need to explain it where to find the mo file. By default, it will
look into /usr/share/locale/de/LC_MESSAGES/messages.mo. You need to
define two things:

1. What is the name of your text domain? (e.g. langtest)
2. What is the location where you put .mo files?

If you use the  gettext style API, you'ld do

gettext.bindtextdomain("langtest","/home/linkweiler")
def _(msg):
  return gettext.dgettext("langtest",msg)

In this case, the mo file should be in
/home/linkweiler/de/LC_MESSAGES/langtest.mo.

If you use gettext.install, you do

gettext.install("langtest","/home/linkweiler")

Again, the file must be called langtest.mo, and reside in
/home/linkweiler/de/LC_MESSAGES.

If you already know the .mo file to use, you can do

_ = gettext.GNUTranslations(open("path-to-mo-file","rb")).gettext

Regards,
Martin