| Age | Commit message (Collapse) | Author |
|---|
|
|
|
This reverts commit 1e8e0da3eb475bb8b78baa54cb0c34b913c2dc5d.
I don't want this going into a release without further review, as well
as documentation and tests, so I'm reverting it (at least temporarily).
See emails on kde-commits mailing list for further rationale.
CCMAIL: notmart@gmail.com
|
|
This technically works and is practiced, e.g. for some Breeze icons.
Some SVG icons do not scale well down from 32 to 22 or up from 16 to 22.
For such cases icons are typically specially crafted for 22 and 16, at least.
Then there's no single icon that be marked as "sc".
So warnings such as:
CMake Warning at ..../ECMInstallIcons.cmake:272 (message):
Fixed-size icon foo.svg is not PNG or MNG
... are misleading.
This change adds SVG(Z) extensions to the list of allowed icons for specific sizes.
REVIEW:126199
|
|
|
|
Because the old implementation (accidentally) worked when you put the
ecm_create_qm_loader call in a different CMakeLists.txt file to the
target the file was added to, some projects did this.
This won't work with build-time-generated files, though, like moc files.
So we (ab)use QTimer events to make the loading happen on the main
thread.
BUG: 346188
REVIEW: 126000
|
|
templates are very useful as teaching tool in order to make
a minimal application that uses a certain framework.
templates in the KAppTemplate repository will always get forgotten
(plus kapptemplate is not really necessary as they work in kdevelop as well)
An ideal situation would be frameworks having templates in their own repos
with templates of barebone apps using the main framework features.
In order to do that, the cmake stuff needed in order to correctly install
a template needs to be ported to a place avaiable to all frameworks
REVIEW:126185
|
|
It should now be easier to read, and more featureful. Among other
tweaks, we now print a summary of dependencies and build options, and
the documentation is generated with more sensible breadcrumbs and
builds properly with Sphinx 1.3.
REVIEW: 126075
|
|
REVIEW: 126090
|
|
The XINPUT library is unlikely to be available on many systems,
including a lot of Linux distributions. This will warn developers in
advance by (a) not searching for XINPUT unless it was explicitly
requested, and (b) printing a warning if they do request it.
REVIEW: 125935
|
|
Makes sure there isn't old stuff still there, it didn't save much time to
keep the files around anyway.
|
|
|
|
Multiple ecm_create_qm_loader() with different catalog names would
overwrite each other's generated files, causing the wrong catalog to be
loaded at runtime for some targets.
This puts the catalog name into the generated filename. Since the
catalog name is the only difference between the generated files, this is
sufficient to fix the runtime behaviour.
REVIEW: 125999
|
|
This is based on commit 6745bd7e4796560959bb67e33b7c7f86f96a5a94 (and
the subsequent fix-up commits).
|
|
This broke the build for projects which used ecm_create_qm_loader in
unusual ways. A better approach is coming, but won't be in e-c-m 5.16.
This reverts commit 009c480413910e8c1a18f4d1420f4a517ea606e6.
CCBUG: 346188
CCMAIL: release-team@kde.org
CCMAIL: kde-buildsystem@kde.org
|
|
This commit is dependent on 009c480413910e8c1a18f4d1420f4a517ea606e6,
which is the primary commit to be reverted.
This reverts commit 634a0a2d3d7ba1bec2bae846acf0faf6184f4c71.
|
|
This commit is dependent on 009c480413910e8c1a18f4d1420f4a517ea606e6,
which is the primary commit to be reverted.
This reverts commit 0496f8ae020c497b5f65234063150781d023f8ba.
|
|
This commit is dependent on 009c480413910e8c1a18f4d1420f4a517ea606e6,
which is the primary commit to be reverted.
This reverts commit 39484722101730d4f1a4bccbc5467c0eb2cc0f6f.
|
|
The V1 syntax of ecm_install_icons searched for icons by globbing files
with a particular naming pattern. If there were no such icons, this used
to do nothing, but silently. Commit fb7b8eea7d accidentally made this an
error. More sensible would be to make it a warning.
BUG: 354610
REVIEW: 125931
|
|
CCMAIL: kde-licensing@kde.org
CCMAIL: agateau@kde.org
|
|
LANGUAGE has higher priority than LC_ALL.
|
|
This is easier to distinguish from "english" in the test output than
"british english".
|
|
|
|
BUG: 346188
REVIEW: 123726
|
|
ECMInstallIcons now updates the theme cache if gtk-update-icon-cache is
available, producing files the test hadn't been expecting.. Updating the
test revealed that the old-style ecm_install_icons call only updated the
hicolor cache, and not any of the other themes.
REVIEW: 125631
|
|
from the normal qt.io installer on mac os
we can revisit this change, if it leads to problems for mac ports or homebrew
REVIEW: 125614
|
|
|
|
Despite the name, Qt is also using this, and it considerably speeds up
icon lookup.
REVIEW: 125192
|
|
* Remove get_property calls on targets, this way we don't need to be called
right before configuration time.
* Removes EOFHook
Instead we process it at generation time using the link.txt file (which is
probably another hack)
REVIEW: 125084
|
|
|
|
Reviewed at https://git.reviewboard.kde.org/r/125163/
|
|
In general, we are looking for what we're installing, not for what we
already had.
|
|
|
|
|
|
QtCore/qsystemdetection.h sets the define Q_OS_ANDROID based on having
ANDROID defined. Hence, adding the ANDROID define allows applications to
use the Q_OS_ANDROID for ifdef'ing.
REVIEW: 125183
|
|
This warning is emitted for every qCDebug() call and is not useful.
REVIEW: 125031
|
|
This adds a new keyword COMMON_HEADER which generates a new header
containing all other headers. Also it is possible now to have multiple
dummy headers per header file. It is assumed that the first header is the
existing one.
REVIEW: 124847
|
|
|
|
|
|
This makes life a bit easier for developers who use the categorised
logging in Qt5 in the common case - rather than creating two new files,
and remembering to put in the #ifdef for the default verbosity settings
in Qt 5.4, they can just add a couple of lines to their CMakeLists.txt.
REVIEW: 124595
|
|
REVIEW: 124763
|
|
(they are slower). And enable them on MSVC now that we rely on Qt 5.3, as the comment said.
REVIEW: 124762
|
|
This prevents accidental "leaking" of symbols from internal code, such
as flex/bison generated parsers.
REVIEW: 124740
|
|
|
|
|
|
In case it fails, offer an error message and the attempted call, to make
sure we can react accordingly.
Note that the proper error argument is FATAL_ERROR, not FATAL.
REVIEW: 124106
|
|
|
|
CI said:
Linking C executable dummy.app/Contents/MacOS/dummy
Built target dummy
Could not find path to executable, perhaps it was not built: dummy
REVIEW: 123969
|
|
This makes it convenient to make further modifications to the tests,
such as setting properties on either the tests or the targets.
CHANGELOG: New arguments for ecm_add_tests().
BUG: 345797
REVIEW: 123841
|
|
ecm_add_test is the less useful (and less used) version of the functions
(because you really might as well just use add_executable() and
add_test() manually in that case).
REVIEW: 123841
|
|
This reverts commit 0c224194ea7f12eaed32af746fc9138537f1919c.
Stephen Kelly pointed out that this is probably not the best approach to
the problem, and runs counter to the direction KDE's CMake code has been
going (splitting functions up and using CMake built-ins where possible).
I have a better solution in mind, which I'll post a review for later.
CCMAIL: kde-buildsystem@kde.org
CCBUG: 345797
|
