Control: severity -1 important
Control: tags -1 + pending
Hi Yuri and Tobias
Thanks for the report and also for looking at additional places where
these paths might be wrong.
We've not successfully managed to get upstream's build system to
understand the split installation approach that we would expect to have
in Debian packages and so there is a reasonable amount of patching or
sed-editing of both the cmake and pkgconf files in the pyside6 build.
Up until now, we also didn't have anyone using it so we didn't know
whether we had managed to get it right... and we had not!
I've added in a couple of autopkgtest tests that are quite trivial in
loading cmake and pkgconf data and then printing out the variables.
- The cmake test will fail for some of the imports if the
files/directories don't exist (as you saw for PYSIDE_TYPESYSTEMS and
PYSIDE_GLUE) although it won't for the INTERFACE_INCLUDE_DIRECTORIES
error that you also spotted.
- The pkgconf test checks that each entry is a file or directory that
exists.
These are pretty weak tests but are a step forward, and I've fixed the
typesystems and glue variables plus the include files for pyside6 (and
also shiboken6 which had the same problem).
I've manually checked the rest of the output of these cmake and pkgconf
variables so I hope that's now correct, but the ultimate test is someone
actually trying to build things using it.
I'd be interested in seeing if we could use one of the PySide6 examples
as more of a test - actually compile the "scriptableapplication" example
for instance. (I'd be very happy to get a patch from someone who is
actually using these C++ bindings -- all current usage of PySide6 in the
archive is from Python, of course).
Please let me know if there any other variables in the cmake files that
aren't yet correct.
thanks
Stuart
--
Stuart Prescott http://www.nanonanonano.net/ [email protected]
Debian Developer http://www.debian.org/ [email protected]
GPG fingerprint 90E2 D2C1 AD14 6A1B 7EBB 891D BBC1 7EBB 1396 F2F7
