Re: New document required

[Telsa Gwynne <hobbit aloss ukuu org uk>]

On Wed, Aug 13, 2003 at 09:10:07PM +0200 or thereabouts, Karl Eichwalder wrote:
> Telsa Gwynne <hobbit@aloss.ukuu.org.uk> writes:
> 
> >    (If anyone has a quick macro or script for "find msgids with
> >    _ marks which have no _ marks in the corresponding msgstr" as
> >    someone recently asked, please speak up!)
> 
> For this job you can use msgfmt:
> 
>       --check-accelerators[=CHAR]  check presence of keyboard accelerators for
>                                 menu items

Blush. Okay, I see what that means now, but I read it and managed
not to realise what it meant. Thanks :)  

> >      - don't put extra [ marks followed by an alphabetical character
> >      in. I don't quite understand this, but msgfmt doesn't like it.
> 
> Works for me:
> 
> echo -e 'msgid "hello"\nmsgstr "h[allo"' | msgfmt - && echo $?
> 0
> 
> Will you please provide an example?

I'm trying to find it now. What was happening was that initially
we were having trouble inserting accents onto letters and some
files contained things like [i"] to mean "turn this into ï" 
(that should be an i with two dots if it didn't come out right).
And I or Dafydd would add the accents in before committing.

And msgfmt whinged about it until it was changed. Of course I
don't remember the error message, and now I can't reproduce it. 
Something along the lines of unrecognised command. I am trying
to get it to happen again, but it won't now!

> >      - You need to count up \ns and put the right number in. If you
> >      don't, msgfmt doesn't like it.
> 
> Yes ;)  Or no ;)  Only the occurence at the beginning or the end of the
> strings is significant.

Ah. I had wondered. Thanks.

> >  * You may also have to dive into (n)gettext arcana if yours is a
> >    language which wants different words for different forms of
> >    plurals. You don't need everyone on the team to do this, but you
...
> Bruno Haible often helps within a day or so:
> 
> ke@tux:~> LANG=C gettext --help | grep bug | sed 's/@/ at /'
> Report bugs to <bug-gnu-gettext at gnu.org>.

Thanks. It's not that it's not working or anything. The problem
is knowing C syntax well enough to know how to describe "here is
the list of numbers which have different plurals". It is not always
as simple as "1" and "more than 1". I know there are loads of
hackers on this list, but it's not guaranteed that every team
will have a handy hacker :) 

Thanks a lot for the comments. I shall go and fix things now :) 

Telsa