[Octave-bug-tracker] [bug #48881] Build failure due to -qt5 flag for uic, rcc, moc, and lrelease

[Dan Sebald]

Follow-up Comment #13, bug #48881 (project octave):

Without reviewing, for brevity, let me just pose this question.  

If the only issue is getting those moc, uic, proxies straight and the -qt5
flag is not something needed for compilation, is the proper route for Octave
configuration to (instead of the -qt5 flag) redefine QT_CHOOSER in the
configure script from

QTCHOOSER='qtchooser'

to

QTCHOOSER='qtchooser -qt=qt5'

or

QTCHOOSER='qtchooser -qt=qt4'

(or "5" and "4")?

I don't have time to investigate right now.  Without the -qt4 or -qt5 flag,
Octave still needs something to make the choice, so the QTCHOOSER needs to
always be redefined (unless we want to leave it up to the user/builder to
modify default.conf, but I don't think that is a good idea).

My thinking is that the user with local build should then define

~/.config/qtchooser/5.conf
~/.config/qtchooser/qt5.conf

and things are good so long as Octave configure recognizes there is a Qt5
version out there.  (And the proper way to do so is probably through
"qtchooser -l", not by searching for Qt libraries.)

And here's another question I have.  Should Octave configure have some option
like

--qtchooser=qt5local

so that QTCHOOSER can be redefined as

QTCHOOSER='qtchooser -qt=qt5local'

in which case the builder needs to place in their account the file

~/.config/qtchooser/qt5local.conf

?


    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?48881>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/