hu.po ngettext fixes

[Egmont Koblinger]

Hi,

Please apply the first attached patch. It fixes the following incorrect
ngettext usage in hu.po:

As the header of hu.po correctly states, as well as mentioned in glibc's
info pages, Hungarian has only one msgstr form, we use the singular form of
a noun if it is preceded by a number.

However, the po file later contains msgstr[0] and msgstr[1] entries, whereas
msgstr[0] (the form that is always being used) is the translation of msgid
(the singular version), and the translation of msgid_plural is never used.

Hence, for example, when running "dd" with the "hu_HU" locale, it always
reports that 1 byte was copied (in Hungarian).

The first two fixes in the patch are the really important ones, the others
are only cosmetical changes since there msgstr[0] and msgstr[1] are
currently the same.


To avoid similar bugs in the future, I highly recommend to change the source
so that it also uses "%d" and stuff in the singular case and substitutes
constant 1 after gettext returns the string. See the second attachment.



Thanks,

Egmont

coreutils-hu-po-ngettext.patch
Description: Text document

coreutils-more-robust-for-translators.patch
Description: Text document