On Sat, 24 Apr 2010 12:10:13 +0200, Andreas Radke <a.radke@arcor.de>
wrote:
> If nothing is broken we can move unixodbc+qt in a few days to extra.

Please don't use versioned deps next time. Also that rebuild was not
needed. Did you use the gcc from testing to build it? If yes, there might
be chance that it cannot be moved without moving gcc-libs too.

I would suggest to completely removed that qt package from testing.

Greetings,

Pierre

--
Pierre Schmitz, https://users.archlinux.de/~pierre

Am Sat, 24 Apr 2010 13:34:11 +0200
schrieb Pierre Schmitz <pierre@archlinux.de>:

> On Sat, 24 Apr 2010 12:10:13 +0200, Andreas Radke <a.radke@arcor.de>
> wrote:
> > If nothing is broken we can move unixodbc+qt in a few days to extra.
>
> Please don't use versioned deps next time. Also that rebuild was not
> needed. Did you use the gcc from testing to build it? If yes, there
> might be chance that it cannot be moved without moving gcc-libs too.
>
> I would suggest to completely removed that qt package from testing.
>
> Greetings,
>
> Pierre
>

The versioned dep is in makedepend and won't break anything for any
user at runtime. The old unixodbc seems to break the build so I think
it's ok this time here. (runtime versioned deps may be questionable in
your view, so I didn't touch your package for runtime deps)

And yes. It's build using gcc because the bug mentioned it broken using
the new gcc. This shouldn't be a problem. I expect the toolchain to
move very soon.

-Andy

On 24/04/10 21:34, Pierre Schmitz wrote:

On Sat, 24 Apr 2010 12:10:13 +0200, Andreas Radke<a.radke@arcor.de>
wrote:

If nothing is broken we can move unixodbc+qt in a few days to extra.


Please don't use versioned deps next time. Also that rebuild was not
needed. Did you use the gcc from testing to build it? If yes, there might
be chance that it cannot be moved without moving gcc-libs too.

I would suggest to completely removed that qt package from testing.


That would result in people who have updated having a newer version of
the package than in the repos.


I am moving the toolchain now so you are free to move that package whenever.

Allan

On Sat, 24 Apr 2010 13:57:35 +0200, Andreas Radke <a.radke@arcor.de>
wrote:
> The versioned dep is in makedepend and won't break anything for any
> user at runtime. The old unixodbc seems to break the build so I think
> it's ok this time here. (runtime versioned deps may be questionable in
> your view, so I didn't touch your package for runtime deps)

Yes, versiond makedepns are not that bad as regular deps, but still
useless and wrong. If you think that's still questionable you should have
respond to my other mail about versioned deps. Maybe I should setup a page
about this anyway to explain things.

> And yes. It's build using gcc because the bug mentioned it broken using
> the new gcc. This shouldn't be a problem. I expect the toolchain to
> move very soon.

Of course, makes sense.

--
Pierre Schmitz, https://users.archlinux.de/~pierre

Now that new gcc is in core OOo also failed over the old unixodbc. So
can we move it together with the rebuilt qt?

-Andy

On Sun, 25 Apr 2010 09:37:40 +0200, Andreas Radke <a.radke@arcor.de>
wrote:
> Now that new gcc is in core OOo also failed over the old unixodbc. So
> can we move it together with the rebuilt qt?

There is not much of a choice. :-) So go ahead.

--
Pierre Schmitz, https://users.archlinux.de/~pierre