tech-pkg archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

upstream math/octave breaks pkgsrc by finding qtchooser



Hi.

I am trying to update math/octave. I stumble over it defining -qt=5 as
argument to the various Qt helpers like moc. This is a thing with
distros where they install qtchooser, see

	https://manpages.debian.org/stretch/qtchooser/qtchooser.1.en.html

for the idea. The octave people added this in

	https://savannah.gnu.org/bugs/?59474

and this breaks the build in pgksrc, as pkgsrc doesn't use qtchooser.

config.log says:

MOC='/data/pkg/qt5/bin/moc'
MOCFLAGS='-qt=5'

But moc says:

$ moc -qt=5 
moc: Unknown option 'qt'.

Well. It's not qtchooser, but /data/pkg/qt5/bin/moc .

I have to correct myself, though. Things break if I have Qt5 installed
in the host system with qtchooser:

configure:77038: checking for x86_64-debian-linux-qtchooser
configure:77073: result: no
configure:77087: checking for qtchooser
configure:77108: found /usr/bin/qtchooser

So if we neuter that check, things should roll again. I wonder if that
will creep up in more Qt packages. The correct test would of course be
to actually check that 'moc' is a symlink to qtchooser. But I guess it
is reasonable of upstream to assume that qtchooser is used if present.

I'll fixup octave now, but wonder. Should a generic fix be applied for
all Qt packages with similar assumptions? Wrappers hiding the qtchooser
command? Or should it be packaged and pkgsrc Qt handling aligned with
what Linux distros do? 


Alrighty then,

Thomas

-- 
Dr. Thomas Orgis
HPC @ Universität Hamburg


Home | Main Index | Thread Index | Old Index