CMake dependency scanning for .moc files

[Alexander Neundorf]

On Wednesday 16 April 2008, Matthias Kretz wrote:
> On Wednesday 16 April 2008, Alexander Neundorf wrote:
> > On Wednesday 16 April 2008, Andreas Pakulat wrote:
> > > On 15.04.08 19:19:45, Alexander Neundorf wrote:
> > > > On Monday 14 April 2008, Andreas Pakulat wrote:
> > > > > Hi,
> > > > >
> > > > > so today I found that there are actually problems with cmake 2.6
> > > > > and dependency scanning.
> > > > >
> > > > > a) after changing an installed header the .moc files for headers
> > > > > that include the updated header are not re-generated which might
> > > > > cause problems when you do ABI changes (like removing a method) on
> > > > > the installed header (linking errors)
> > > > >
> > > > > b) removing a .moc file from the builddir doesn't produce a re-moc,
> > > > > cmake just tells gcc to compile the non-existing .moc-file.
> > > > >
> > > > > I'm actually not sure wether either of the two ever worked, but
> > > > > IMHO at least the second one should work.
> > > > >
> > > > > This is was tested on CMake 2.6 and 2.4.7.
> > > >
> > > > And it works with 2.4.7 but doesn't with 2.6 ?
> > >
> > > It doesn't work with either of the two.
> >
> > Ok, so it's at least no cmake 2.6 problem, but automoc.
> >
> > > > This is a moc file handled via automoc, right ?
> > >
> > > Yes. The only way to get cmake to re-generate the .moc file is to touch
> > > the header from which its generated or remove the subdir in the
> > > builddir that contains the .moc file.
>
> In the resulting Makefiles there are no dependencies on the actual moc
> files. The upside of this is that changing a moc-include doesn't need a
> cmake rerun (that includes adding Q_OBJECT to a class - before kde4automoc
> you had to rerun cmake then). Even if the Makefile doesn't have a dep on
> the moc files, if kde4automoc is run it will compare the timestamps of the
> moc file against it's source file (header or cpp), that's why touching the
> .h file works.
>
> IIUC you want to look at the timestamps of all the headers that are
> directly or indirectly included from the file that is moced. AFAIK not even
> cmake generated Makefiles do that for compilation (I imagine that would be
> rather expensive information). If they do I'd like to know how to get at
> that dependency info to reuse it for kde4automoc.
>
> But really, I think that if the installed headers change in a binary
> incompatible way the only safe thing you can do is a make clean.
>
> Conclusion: I think everything works as it should. At least from the
> problem description above.

For case a) I agree, but what about the deleted moc file in case b) ?

Alex