Review Request 118264: Switch to PolkitQt5-1

[Aleix Pol Gonzalez]

> On May 28, 2014, 12:56 p.m., Aurélien Gâteau wrote:
> > Building with polkit-qt-1 master (bac771e69887c9253f2b0973f6310810db0061f8) fails with this error:
> > 
> > [ 55%] Building CXX object autotests/CMakeFiles/KAuthHelperTest.dir/HelperTest.cpp.o
> > In file included from /home/aurelien/build/kf5/frameworks/kauth/src/moc_Polkit1Backend.cpp:9:0,
> >                  from /home/aurelien/build/kf5/frameworks/kauth/src/kauth_backend_plugin_automoc.cpp:3:
> > /home/aurelien/build/kf5/frameworks/kauth/src/../../../../../src/kf5/frameworks/kauth/src/backends/polkit-1/Polkit1Backend.h:31:31: fatal error: PolkitQt1/Authority: No such file or directory
> >  #include <PolkitQt1/Authority>
> >                                ^
> > compilation terminated.
> > In file included from /home/aurelien/src/kf5/frameworks/kauth/src/backends/polkit-1/Polkit1Backend.cpp:22:0:
> > /home/aurelien/src/kf5/frameworks/kauth/src/backends/polkit-1/Polkit1Backend.h:31:31: fatal error: PolkitQt1/Authority: No such file or directory
> >  #include <PolkitQt1/Authority>
> >                                ^
> > compilation terminated.
> 
> Hrvoje Senjan wrote:
>     >Building with polkit-qt-1 master (bac771e69887c9253f2b0973f6310810db0061f8) fails with this error:
>     
>     you're right. bac771e in polkit-qt-1 broke it. i've tested with official tarball + commit ddca404 ('Add Qt 5 support').
>     with polkit-qt-1 buildsystem that commit seems wrong. Kevin, can you tell what was reasoning for the change?
> 
> Kevin Ottens wrote:
>     When I did that commit, @INCLUDE_INSTALL_DIR@ and @LIB_INSTALL_DIR@ were absolute path. Might have changed in the meantime of course.
> 
> Hrvoje Senjan wrote:
>     with that commit, and without passing neither INCLUDE_INSTALL_DIR or LIB_INSTALL_DIR, one gets:
>     set(POLKITQT-1_INCLUDE_DIR              "include/polkit-qt5-1")
>     ...
>     set(POLKITQT-1_LIB_DIR                  "lib64")
>     
>     so, the commit fixes the vars when they are absolute, but breaks when relative paths are used, and when no special path is used.
>     just to note - it was like this before, the 'Qt5 port patch' temporary 'fixed' it for the above two cases.
>     i guess it needs more bulletproof solution.

Well, I would say PolKit-Qt needs to be ported to exporting the targets and the include directories with them.
Hrvoje, do you want me to take a look at it? Or you're planning to?

Either way, I don't think we should push that patch if it's not going to compile anyway, given that today is the tagging.


- Aleix


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://git.reviewboard.kde.org/r/118264/#review58664
-----------------------------------------------------------


On May 28, 2014, 2:01 p.m., Hrvoje Senjan wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://git.reviewboard.kde.org/r/118264/
> -----------------------------------------------------------
> 
> (Updated May 28, 2014, 2:01 p.m.)
> 
> 
> Review request for KDE Frameworks and Kevin Ottens.
> 
> 
> Repository: kauth
> 
> 
> Description
> -------
> 
> since now master of polkit-qt-1 is able to exist w/o issues with Qt4 variant, KAuth should switch to that, instead of building either w/ no backend, or with the branch...
> 
> 
> Diffs
> -----
> 
>   cmake/KF5AuthMacros.cmake 4cdf3ab 
>   src/ConfigureChecks.cmake aff05ed 
> 
> Diff: https://git.reviewboard.kde.org/r/118264/diff/
> 
> 
> Testing
> -------
> 
> Got CMake message:
> Building PolkitQt5-1 KAuth backend
> and kauth-policy-gen also built
> KAuth macros generate correct polkit-1 rules (e.g. date kcm in plasma-desktop, etc)
> 
> 
> Thanks,
> 
> Hrvoje Senjan
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kde-frameworks-devel/attachments/20140630/e5a6c9de/attachment-0001.html>