On 8/20/23 16:42, Thomas Baumgart via KMyMoney-devel wrote:
On Sonntag, 20. August 2023 20:13:27 CEST Jack via KMyMoney-devel wrote:
It seems the problem with cmake not finding webenginewidgets has been
popping up in a number of different threads in different support
venues. In my current case, trying to compile from git head master
branch gives me
CMake Error at
/usr/lib64/cmake/LibAlkimia5-8.1/LibAlkimia5Targets.cmake:65
(set_target_properties):
The link interface of target "Alkimia::alkimia" contains:
Qt5::WebEngineWidgets
but the target was not found. Possible reasons include:
This is on artixlinux, where webenginewidgets is provided by the
qt5-webengine package. The relevant cmake files in that package are
qt5-webengine usr/lib/cmake/Qt5WebEngineWidgets/
qt5-webengine
usr/lib/cmake/Qt5WebEngineWidgets/Qt5WebEngineWidgetsConfig.cmake
qt5-webengine
usr/lib/cmake/Qt5WebEngineWidgets/Qt5WebEngineWidgetsConfigVersion.cmake
Note on my system /usr/lib and /usr/lib64 are symlinked. However, I
wonder if the call should be for Qt5::Qt5WebEngineWidgets.
I am not sure at all. I have the following files here
/usr/lib64/cmake/Qt5Designer/Qt5Designer_QWebEngineViewPlugin.cmake
/usr/lib64/cmake/Qt5WebEngine/Qt5WebEngineConfig.cmake
/usr/lib64/cmake/Qt5WebEngine/Qt5WebEngineConfigVersion.cmake
/usr/lib64/cmake/Qt5WebEngineCore/Qt5WebEngineCoreConfig.cmake
/usr/lib64/cmake/Qt5WebEngineCore/Qt5WebEngineCoreConfigVersion.cmake
/usr/lib64/cmake/Qt5WebEngineWidgets/Qt5WebEngineWidgetsConfig.cmake
/usr/lib64/cmake/Qt5WebEngineWidgets/Qt5WebEngineWidgetsConfigVersion.cmake
which looks very similar to yours and have no problems. I do use
alkimia master though, not the 8.1 branch.
Is there a terminology difference (qt5webenginewidgets vs
webenginewidets) between different packages providing and using this
component? If so, which is right and which needs to be corrected? Or
am I totally off the mark and this difference is not relevant?
So am I. I don't know if that is caused by naming conventions of
the various distros or something else? Also, why is it failing now
and did work in the past?
I also just compiled 8.1.1 from tarball, and it produced the same
files. However, if running cmake on the tarball of libalkimia found the
necessary files, then why can't they be found when the search is
transitive, initiated by compiling KMM from git master? Doesn't 5.1 call
for the same files when using webengine instead of webkit?
What I now wonder is whether something changed in the cmake files in
KMyMoney master branch which alters something about the search path,
especially with a nested cmake file. My next effort will be capturing
the logs of cmake for both 5.1 and master to see if I can find any
relevant differences.
Curiouser and curiouser, said Alice.
