• change BEFORE to BEFORE_AND_AT, to match actual Qt behaviour (though do not copy confusing name, but be more precise using AND_AT
• add NO_DEFINITION_EXPORT_TO_BUILD_INTERFACE option

In D23789#527769, @cgiboudeaux wrote:

Tests needed :)

Tests needed :)

Implementation prototype for T11490.

Given ECM 5.62 will be tagged this WE and thus the API of the new macro ecm_qtdesignerplugin_widget being frozen, and no-one really gave feedback on the initial API, would take the responsibility and push tonight, unless there is last minute discussion. Only users of the arguments changed here is Okteta (by what lxr.kde.org reports), which are my changes as well and partial motivation for this change.

For convenience still enable CONSTRUCTOR_ARGS_CODE, usually simple string
without ";"

[spam comment removed by sysadmin]

[spam comment removed by sysadmin]

Ah, I didn't realize there was no warning if calling set_package_properties with different properties.

I'd do the opposite, remove DESCRIPTION and URL from knotification's CMakeLists.txt and only leave the PURPOSE line.

Compare though all the

Ideally this is set already directly in the Find-module.

for DESCRIPTION and URL properties in the docs https://cmake.org/cmake/help/latest/module/FeatureSummary.html

No objections from me. This was explicitly suggested during review though, so I'd like @cgiboudeaux to approve this.

@heikobecker ping!

KF6 is a long time out so I'm abandoning this.

KF6 is a long time out so I'm abandoning this.

https://gitlab.kitware.com/cmake/cmake/issues/19615

Thanks!

D23262

Yep, KTextEditor is broken for me, too.

This seems to cause the build of sonnet to fail: https://invent.kde.org/snippets/394

LGTM

typo--

bump version again

Thanks!

Patch as is fine with me.
But no expert on both flags itself, no idea if there could be sideeffects of having those two now injected as ON to all projects. Possibly best to do global rebuilds of all projects on CI once this lands.

I'm happy with this, +2 from my side.

@kossebau @vkrause @apol something else to change?

This shall be compliant with the minimum ECM version set in the project.

Marco explained privately he doesn't want this so I'll move it to my codebase for now until the alternative better thing appears (maybe I'll try to make it appear soon).

In D22958#507299, @nicolasfella wrote:

Why do you restrict it to SVGs? Wouldn't you be able to use PNGs as well?

Why do you restrict it to SVGs? Wouldn't you be able to use PNGs as well?

The patch makes sense. +1

This is the culmination of a patch series that makes About dialogs in Kirigami work reasonably across platforms:

The code is mostly copy-and-pasted together from ECMAddAppIcon.cmake and ECMInstallIcons.cmake.

Ping

Ping

Seems people so far are fine with this, any comments were rather positive. No changes requested.

document behavior and when it was introduced, also enable autouic

Indeed, both good points.

+1, same for AUTOUIC probably?

Just a small principal comment with "supporting-old-versions" hat on.
Cannot give feedback on autorcc itself, never used, so no idea how useful/reliable it is.

+1 I've been using it in several project, it works good. I don't see how this could be misused.

+1 LGTM

• escape strings to be used as strings in C++ code
• use Qt5Core version to decide whether checking only for QtUiPlugin is enough

add more options: DEFAULT_HEADER_CASE, DEFAULT_HEADER_EXTENSION, DEFAULT_ICON_DIR

Accidentally uploaded the wrong diff.

• more code comments
• fix icon file existance check

• API DOX: explain arguments
• add argument to make it work for Okteta's extra code
• avoid rebuilding if generated file did not change
• place generated files in subdir for namespacing

Update CMake minimum required version to 3.1 (for target_sources).

I

I'm not sure if its beneficial to use cmake_parse_arguments here. It seems like it would make the macro harder to understand compared to the current change. I'm also not entirely sure how it would work? I haven't worked with cmake_parse_arguments before.

How about pass it as a TARGET argument using CMakeParseArguments.
You can see documentation about it here: https://cmake.org/cmake/help/latest/command/cmake_parse_arguments.html

LGTM

address cgiboudeaux's comments