Re: [Pingus-Devel] Clanlib 0.8 has been released

[Ingo Ruhnke]

[This mail is a repost, since first one seems to have disappeared
somewhere]

Hans de Goede <address@hidden> writes:

>>> Yes, that's what I'm doing. Grumble shouldn't be dead AFAIK just
>>> not active on Pingus anymore. If need be he should be reachable.
>>
>> I kicked gettext out for a reason,
> And that reason was if I may ask?

Because its inflexible, unportable and quite limited in its API, ie.
no proper way to browse which languages are available, no proper way
to extent dictonaries (custom levels might provide their own .po),
requirement to compile .po, no way to run directly from source
directory and other stuff.

> Whats wrong with it is that it cannot handle character encoding
> conversion in any way (at all)

It converts from the charset in the .po to the one used by the bitmap
font. That won't do much good if one tries to output to the terminal,
which is certainly a limitation, but I don't think gettext solves that
any better.

> and that it doesn't searches for translations on unix in the usual
> way / location,

Thats a feature, since it allows to run both without compiling the .po
as well as to use it from whereever the data files are stuck, which
gettext can't.

> breaking packaging tools like rpm, which expect this location and
> use it this "info" to only install the languages requested by the
> user.

Bad luck, but that seems to be like a problem of the distribution not
pingus. If the distribution can't handle that, it won't remove the
unneeded fonts from Pingus either, which waste a more disk space then
thoes few .po files anyway.

> Also if you would have bothered to take a look you would have seen
> that the patch is very non intrusive and only uses gettext if
> HAVE_GETTEXT get defined, if that is not defined then the current
> behaviour and code paths are kept for 100%.

I don't care how non-intrusive it is, its still code I don't care
about and which will surly break stuff in the future, for one because
I simply don't care to test and secondly because it doesn't provide
the feature I will need. Remember we are talking about an completly
unfinished SVN snapshot here, that in no way shows the final features
of an official next Pingus release.

> Again if you would have bothered to take a look you would have known
> that it actually fixes something, I'm not in the habit of comitting
> useless patches.

 From a quick look it seems like it moves the init_path_finder()
further down the line, which would cause --help output to go out
untranslated, which of course is kind of a tricky issue, since
--datadir gets set at a point where console output might already have
happened.

> but maybe its time to pass pingus leadership over to someone else?

If somebody wants that job sure, absolutly no problem, but so far
nobody has volunteered.

-- 
WWW:      http://pingus.seul.org/~grumbel/
Blog:     http://grumbel.blogspot.com/
JabberID: address@hidden 
ICQ:      59461927