because cursors are technically images and the thumbnails creation is cheap
(even tho the result is not cached, cursors are small images)
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
animating icons was never a KIconLoader thing, animated icons however (such
as "process-working", which is many icons packed into one) are supported.
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
the systemtray applet automatically adds the applets that have
configuration entry in plasma-desktoprc but the applets configuration is
not saved by the systemtray applet
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
for reference:
https://github.com/fluxer/katana/issues/41
note that the config affects only new configurations (the config is
copied to the user config directory once layout is saved), also a logout
may be required when a screen is removed (that includes setting up extra
monitor to be copy of the primary display)
as for why it is done - a image and panel containment was created by the
default layout script (00-defaultLayout.js) which is no longer since
772d9ff219
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
KLocale::formatDuration() does pretty much the same as what
KLocale::prettyFormatDuration() used to do
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
as the comment says - there is no interface for it, there isn't one for
what KGlobalSettings::isMultiHead() checks for but it checks for
environment variable. now the klauncher KDE_MULTIHEAD environment variable
will match the global envionment variable
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
not only unused but next to useless - it is about a configuration change
and the common way for that to happen is doing it from the randr KCM
which is a "I know what I am doing, I don't need popup for that" kind of
situtation
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
many plasma runner have this issue - when locale is changed to something
other than the default the trigger term has to be in translated form (e.g.
"отметки") because the original untranslated term is not matched (e.g.
"bookmarks") and it is as if the runners are not working
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
on a side note generating backtrace via drkonqi was not working on some
systems where traces are disabled by default (for security reasons) and
it required not only to install debug packages or build for debugging
purposes but also to setup the system so that backtrace can be generated
to get something out of it
where that is no the case on Linux there is usually a core dump in the
system log anyway thus the only thing a program related to crashes has
to do is to notify about it and (when possible) automatically file some
information (with the option to disable that)
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
event tho TaskManager::TaskManager is singleton manually removing the
catalog can result in empty strings (catalogs are no longer reference
counted)
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
std::vector<T> for when the data does not have to be shared (e.g. POD
types) and there is no Q_DECLARE_TYPEINFO() involved, QVector<T> for when
data should be shared
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
it is essentially container for other modules and all but the randr
module are not valid (not maintained as part of Katana)
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
