CMake 4 upload to unstable

[Adrian Bunk]

On Wed, Dec 03, 2025 at 01:29:50PM +0100, Emilio Pozuelo Monfort wrote:
> On 02/12/2025 07:49, Adrian Bunk wrote:
> > On Thu, Sep 25, 2025 at 11:29:08AM +0200, Emilio Pozuelo Monfort wrote:
> > > On 25/09/2025 11:21, Timo Röhling wrote:
> > > > Hi,
> > > > 
> > > > * Emilio Pozuelo Monfort <pochu at debian.org> [2025-09-25 10:07]:
> > > > > Yesterday when I read Simon's email I was going to suggest the same.
> > > > > Raise the severity now, and wait a bit more for packages to be fixed,
> > > > > as otherwise this may cause issues to ongoing or planned transitions.
> > > > That sounds like a reasonable approach.
> > > > 
> > > > > Maybe we can evaluate it again in one month, and hopefully get it
> > > > > uploaded to sid soon.
> > > > I am a bit unhappy about the fuzziness of the "hopefully soon" part, and
> > > > I would prefer something more concrete that the Release Team is
> > > > expecting to happen / where the priorities are. For instance, am I
> > > > correct to infer from Paul's mail that dealing with key packages is more
> > > > important than merely reducing the overall number of open bugs fast?
> > > 
> > > Yes, key packages are more important in general, as it's harder to get rid
> > > of those in testing if there's a need to unblock a transition.
> > > 
> > > I think we can do it in one month if things look reasonably well, otherwise
> > > in two months as a hard deadline to not delay this indefinitely. Does that
> > > sound reasonable?
> > 
> > This was over 2 months ago, and the number of key packages that FTBFS
> > with CMake 4 is now lower than the number that do still FTBFS with GCC 15.
> 
> The llvm bugs worry me a little bit. Other than that, this would be good to go.

#1113237 contains plenty of investigation from Timo, with the open 
question for the LLVM maintainers who know how their debian/rules 
machinery works.

> Cheers,
> Emilio

cu
Adrian