Re: RFS: warsow (updated packages)
* Jon Dowland <lists@alcopop.org> [2007-09-07 13:15:17 CEST]:
> The first thing I notice is a fair number of build warnings along the
> lines of
>
> game/q_math.c:910: warning: passing argument 1 of ‘Matrix_Copy’
> from incompatible pointer type
>
> That example was because the prototype specifies a const argument and is
> passed a non-const argument. (argh, I've managed to avoid looking at C++
> for so long!) so I suppose these are pretty benign. They could be fixed
> with a cast in the caller I think. Would upstream accept a patch to that
> effect? (this would just be nice, I wouldn't think this was a reason to
> delay the package)
Yeah, build warnings are pretty regularly ignored throughout the whole
archive, but looking at them is a good thing from time to time
(including doing patches, yes). And I would highly hope that upstream
accepts them, they can easily become real problems, it wouldn't be the
first thing that got stricter in gcc and turned from a warning into an
real error.
> On Fri, Sep 07, 2007 at 11:03:07AM +0200, Gerfried Fuchs
> wrote:
> > I wondered about the need for repacking the source. I've
> > read README.Debian and to be honest, given that the
> > package doesn't move to main due to removing the things I
> > don't see the real benefit of doing a dfsg version of the
> > source package.
>
> I can't find README.Debian in the source package, nor the generated
> binary package. Am I doing something stupid?
To be honest, I was taking a look at the README.Debian from my
installed warsow package, version 0.31.dfsg-6. If that was removed from
a future version I would consider it a bug because it's useful
documentation.
#v+
r4067 | ceros-guest | 2007-09-05 02:48:12 +0200 (Wed, 05 Sep 2007) | 1 line
Changed paths:
D /packages/trunk/warsow/debian/README.Debian
...
Checking in new packaging for warsow-0.32
#v-
No idea why it was deleted, maybe just on accident. ceros?
[config.{guess,sub} and lintian]
> I like to handle this with a Build-Dependency on
> autotools-dev and ln/rm calls in the right places to symlink
> that package's versions of those two files. What do you
> think of that approach?
Sounds much more reasonable than to removing it from the source package
to avoid a lintian error...
So long,
Rhonda
Reply to: