[bug-gettext] [bug #48411] xgettext does not detect conflicting plural forms

[Guido Flohr]

URL:
  <http://savannah.gnu.org/bugs/?48411>

                 Summary: xgettext does not detect conflicting plural forms
                 Project: GNU gettext
            Submitted by: gflohr
            Submitted on: Tue 05 Jul 2016 02:32:44 PM GMT
                Category: None
                Severity: 3 - Normal
              Item Group: None
                  Status: None
                 Privacy: Public
             Assigned to: None
             Open/Closed: Open
         Discussion Lock: Any

    _______________________________________________________

Details:

Run this piece of C code through xgettext:

  printf(ngettext("One network.\n", "%s networks.\n", count), "Many");
  printf(ngettext("One network.\n", "%lu networks.\n", count), count);

You will get one single po entry:

  #: network.c:1 network.c:2
  #, c-format
  msgid "One network.\n"
  msgid_plural "%s networks.\n"
  msgstr[0] ""
  msgstr[1] ""

In the example code this is likely to cause a segmentation fault.  Yes, sure,
the example does not make a lot of sense but it is perfectly legal usage of
ngettext and illustrates the problem.

The best solution would probably be to use the combination of the singular and
plural message id as the key into the translation catalog.  But that would not
be possible without changing the gmo format.

But xgettext should at least treat this situation as a fatal error condition.

In case somebody runs into this problem in a real-world application: Use
npgettext() instead of ngettext() and add a message context as a workaround!





    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?48411>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/