Conversation
|
dont forget to change line 5 in ports/qt5-base/vcpkg-cmake-wrapper.cmake to link the debug libraries |
|
hit a problem on qt5-tools not using the correct path for a debug library, will look into this tomorrow |
@TheScarfix: let me have a wild guess without looking into it: libpq. |
|
@Neumann-A no, some qtqml library which is searched in the lib path not the debug/lib |
|
still hitting an error on qt5-tools Error: dependent 'C:\dev\vcpkg\installed\x64-windows\lib\Qt5QmlDevToolsd.lib' does not exist. |
|
does It also looks like fixcmake.py is only correcting paths with forward slashes? Maybe that is the issue? |
|
don't think qt building uses cmake in any way. Qts build config says it looks for libraries in debug/lib 🤔 |
|
what I would do in these circumstances is a search all for |
|
So the next questions would be:
|
|
theoretically it looks like |
|
~~hmm ~~
edit: dont do that. Qt5Helpd is build by qt5-tools |
|
After thinking about it. The problem may be in all files in It would be interesting to stop cmake right after/before #6127 (comment) and see if the offending line is already written to the makefile or not. |
|
Something is strange.... |
|
it is really weird because it's the only library with problems 😕 |
|
looked at it with procmon and it looks like qmake rewrites the file wrong. Can't see why though |
|
Its probably because of what is standing in the mkspecs for the different modules in |
|
does d299ad6 solve the problem? |
|
no, still failing |
|
I changed line 4 in qt_lib_qmldevtools_private.pri to: with the following result: I also had merged in latest vpckg master and pr #6259 |
|
could be that it should be |
|
yeah works also with so there are two questions: should we simply set QT_MODULE_HOST_LIB_BASE to QT_MODULE_LIB_BASE in qt_lib_qmldevtools_private.pri? Probably yes after the installation of qt5-base or declarative? (Who creates the file?) |
|
will you make a pr to my branch like last time? |
|
@TheScarfix made a pr. You must test the rest. Compiling qt with nmake always takes forever... (jom stalls on my pc for some unidentified reason) |
fix qmldevtools debug linkage for qt5-tools
|
If you resolve the merge conflicts with latest master then the CI system should pick it up. |
|
This is what I see in the CI system now:
|
|
what do the logs say for those? |
|
@Rastaban: Cannot reproduce x86-windows regressions. Will try x64-windows after full build of qt on x86. |
|
With the latest merge the failures have gone away. To me this PR looks good, is there anything else that needs to be done before we merge? |
|
I think it's good now :) |



closes #5929