Re: Compile problem on multiple non-amd64 architectures
Hi,
I am still not having any luck. I did add "export QT_SELECT=qt5" in my
Makefile (which is directly called by debian/rules from the clean/build
targets) but the error remains the same.
To reproduce the problem, I have logged into a porterbox (mipsel, which
failed on a buildd). However, for me packagesearch compiled fine.
Could it make a difference, whether I export QT_SELECT in the Makefile
or in debian/rules. Since I cannot reproduce the build failure, I am a
bit left in the dark. I hate to upload a lot of new versions to
unstable just to try things out.
Any ideas on how to reproduce the bug? What could be special on the
buildds?
Regards
Ben
On Wed, 2017-11-01 at 16:44 -0300, Lisandro Damián Nicanor Pérez Meyer
wrote:
> On 1 November 2017 at 15:04, Benjamin Mesing <bensmail@gmx.net> wrote:
> > Hi Lisandro,
> >
> > thanks for the quick help. I will try this. Everything looked fine to
> > me, because the qt5-moc was used. So, if I understand the fix
> > correctly, I have to tell the qt5-moc to actually expect qt5 code.
> >
> > It is a little bit confusing, all my makefile are generated from .pro
> > files with "qmake -qt5 input.pro", so I had expected the Makefiles to
> > be generated correctly (i.e. calling moc -qt5). Obviously I was wrong.
> > I wonder, if this is, how the tool should behave?
>
> qtchooser is not mandatory by upstream. Check qtchooser's man page and
> the URL I gave you before, it's better to just set QT_SELECT in
> debian/rules.
Reply to: