Hebrew: update

[Frédéric Delanoy]

2011/8/27 Francois Gouget <fgouget at free.fr>:
> On Sat, 27 Aug 2011, Frédéric Delanoy wrote:
>
>> On Fri, Aug 26, 2011 at 15:14, Francois Gouget <fgouget at free.fr> wrote:
>> > On Tue, 23 Aug 2011, Yaron Shahrabani wrote:
>> > ...
>> > On the translation side I still think that copying untranslated strings
>> > over is wrong <...>
>>
>> Actually, for standardization, empty msgstr should also be marked with
>> ", fuzzy" IMHO.
>
> Why?
>
> An empty msgstr is unambiguously untranslated. Also I would define the
> standard by what the gettext tools do and when they generate a new po
> file they don't mark all empty msgstrs as fuzzy.

OK didn't know that
> If the problem is that you have trouble finding untranslated messages in
> a text editor, then that's one point where having a dedicated PO editor
> helps (or emacs' PO mode).

Not really a problem, it's just easier to find when you can have
translated multiline msgstr like
msgid "some long text"
msgstr ""
"some very long"
"translation"

(I have to use relatively complex regexp to find those cases).

To remain consistent with the gettext tools, for multiline
translation, the generated po entry should be IMO sthg like
msgid "some long text"
msgstr "some very "
"long translation"

i.e. the start of the translation remains on the same line as 'msgstr'

Frédéric