In D13148#269205, @davidedmundson wrote:

This doesn't achieve anything useful.

It does for projects that have their own build system.

You're not changing any dependencies as you still link against DBus as CMake is unchanged in this patch.

I understand this does not change how the upstream build is done, but allows users to keep depending only on Qt.

I did not patch CMakeLists.txt because I can't test it. I could not get a system with all the right versions of all build dependencies to build Kirigami.

On Subsurface-mobile we simply have our own integrated qmake/cmake files that builds Kirigami statically with the rest of the project.

Code behaves the same as not having KWin at runtime will result in both values being the exact same values as we're doing with this patch

So the real issue here is the new dependency on linux. Whatever way we can use to disable it is fine.

In D13148#269233, @murillobernardes wrote:

I did not patch CMakeLists.txt because I can't test it. I could not get a system with all the right versions of all build dependencies to build Kirigami.

On Subsurface-mobile we simply have our own integrated qmake/cmake files that builds Kirigami statically with the rest of the project.

yeah, i know how is done... tough is it a build time or a runtime problem? there is an environment variable to force tablet mode alreeady

Code behaves the same as not having KWin at runtime will result in both values being the exact same values as we're doing with this patch

So the real issue here is the new dependency on linux. Whatever way we can use to disable it is fine.

what would make sense, instead a force_tablet_mode , is to make the dependency to dbus optional on all platforms

In D13148#269556, @mart wrote:

In D13148#269233, @murillobernardes wrote:

I did not patch CMakeLists.txt because I can't test it. I could not get a system with all the right versions of all build dependencies to build Kirigami.

On Subsurface-mobile we simply have our own integrated qmake/cmake files that builds Kirigami statically with the rest of the project.

yeah, i know how is done... tough is it a build time or a runtime problem? there is an environment variable to force tablet mode alreeady

Build time. No tabletmodemanager_interface.h available.

Code behaves the same as not having KWin at runtime will result in both values being the exact same values as we're doing with this patch

So the real issue here is the new dependency on linux. Whatever way we can use to disable it is fine.

what would make sense, instead a force_tablet_mode , is to make the dependency to dbus optional on all platforms

Agreed. What would be an acceptable way to do this then?

almost there :)
since there is also a qmake mode to build for embedded, where dbus should be disabled as well, I would prefer to flip the logic.
so, have a KIRIGAMI_ENABLE_DBUS which is undefined by default (so nothing to do in the .pri file)
but having it defined by default in the main cmake file, so would become

NOT(APPLE) AND NOT(DISABLE_DBUS))
​ find_package(Qt5DBus)
​ add_definitions(-DKIRIGAMI_ENABLE_DBUS)
endif()

If we have an defined(KIRIGAMI_USE_DBUS) we don't need to put
if (!APPLE && ......) in the code which would be much much cleaner.

ping?

Sorry, first time contributing to kde-related project.

I’m not sure what I need to do next? Send a pull request somewhere?

In D13148#271691, @murillobernardes wrote:

Sorry, first time contributing to kde-related project.

I’m not sure what I need to do next? Send a pull request somewhere?

No, you just would land the patch. I'm guessing you don't have contributor powers as it's your first time, I'll land it for you.