in full Katana session it is done by the input KCM and even by the
`startkde` script (so that cursor theme applies for any error messages
shown via `xmessage`). as for changing after startup it is again done by
the input KCM (via `XFixesChangeCursorByName()` calls and other stuff), the
change even applies when not in Katana session (i.e. `systemsettings`
changes the cursor theme even if not in Katana session) so it is redundant
to do it from KGlobalSettings and only the notification signal has to be
emitted
on a side note style and palette are applied by the platform plugin (see
kde-workspace/qguiplatformplugin_kde/qguiplatformplugin_kde.cpp) when in
full Katana session meaning style and palette also do not have to
applied by KGlobalSettings but that is for another commit because there
are some other things to take in account (such as style override via
command line argument and initializing plugin paths early, before
KApplication is constructed)
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
the input KCM (which changes mouse settings) emits that change so it makes
sense to apply mouse-related settings to the application when it is
emitted. also notfication for KGlobalSettings::SETTINGS_QT was never
emitted
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
just another way to do what ASN does, the KService::DBusWait mode was
not used too. with this change however all of the process setup code is
moved to a seperate class and the responsibility of KLauncher about ASN
is reduced (ASN now works better for process that fork but if application
claims ASN support and does not send ASN finish then the timeout will be
reached)
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
fixes the "waitforwm" command line option
(KApplicationPrivate::parseCommandLine() was called before the X11 atom was
created and atom was invalid) and doubles as optimization
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
I intend to rewrite kconfig_compiler at some point (possible to use JSON
format) and the tests will be done anew
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
is interesting how much simpler it can be by assuming that the tags are
strict and by using Qt::mightBeRichText() for the rich text detection.
while mostly compatible (entities are not converted, some tags were not
used and some even did not do anything) it is subject to optimization by
using QStringMatcher or other tricks (like using a separate tag for the
numbers precision)
overall - no regular expression matching, no XML parser is used and does
what the old implementation was doing but with less code. also passes
most tests (the exception is KLocalizedString test case that expect
entities to be replaced)
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
there were 3 copies of it, only one with minor changes. by moving it
here any changes that have to be done to it will not have to be done to
the copies aswell.
TODO: move its private members to KTimerDialogPrivate
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
this makes KIcon serializable by name (e.g. by KDBusMenuExporter) in
most cases, that however means lookup on the other end (if the icon pixmap
is not serialized it is essentially the same as calling QIcon::fromTheme()
on the other end of the serialization)
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
KRatingPainter and KRatingWidget are currently not used but holding a
reference to a pixmap that can already be (and is) obtained from the icon
(constructing QIcon from pixmap is a thing) is just redundant
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
this is mostly for status notifier dialogs, if the dialogs are not raised
and activated they will block (because QDialog::exec() is called) and
manual activation has to be done (via alt+tab)
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
does nothing but slow down the image serialization on little endian really,
the image data is not ment to be saved and passed around to other host
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
Katie has the option to set printer properties and options anyway, as
for why the bits are non-operational - it is because of the version
check which is lower than the current Katie version
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
the homepage and the bugs email address shall always be defined in
kglobalsettings header for official Katana applications, libraries, etc.
the option to override them is still a thing
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
QDesktopWidget is Xinerama aware and it can be disabled because I
implemented option for that, see the following commit in the Katie repo:
b7e4bae65f0cae7f3029b98980f638c077ecfc22
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
because it returns only one (definitive) format type using QStringList as
return type makes no sense, also a mode argument is very much required to
return correct result - not all image format plugins support writing
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
much like KPasswordDialog which does a lot of focus tricks depending on
its flags and which widgets are to be shown
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
to avoid conversion from QString to QByteArray when possible, the trace is
handled as array of latin1-encoded strings anyway
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
a crashed program trying to restart itself is very unreliable,
especially program that uses KUniqueApplication instance. so, instead
all the information about the crash shall be written to a file on the
disk (in the temporary directory) and read by external kcrash KDED
module that will report the crash (possibily even upload the crash
details somewhere) and restart the program as necessarry.
this also opens up possibility for per-application configuration for
things such as the automatic restart feature but that shall be done by the
kcrash KDED module instead (parsing a config from a crashed program is
not a good idea)
KCrash::NoRestart will be unused but kept for future expansion
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
the plasma frame applet for example uses zero time (QTime(0, 0, 0), not
invalid time) as a special for case for "never" (i.e. don't update) so
having this feature will make KTimeEdit even more suitable for timers (that
ofcourse has to be handled by the application or library, zero time timers
are a thing)
also min/max now allows to set lower seconds (for example) when the minute
is not zero instead of just restricting the seconds value to a hard-minumum
altough that makes me think a step feature may be needed (e.g. to now allow
intervals less than two seconds with minimum of one second)
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
