after changes to multiple sub-projects (Katie, kdelibs, kde-workspace
and kde-extraapps) that finally happened, can write lots about it but I
will keep it short - by rewriting parts that were using js/qml (e.g. the
plasma applets) the desktop startup was reduced to half, also the legacy
tray thing shall be no more (anything that needs tray icon shall be a
plasma applet now) thus some applications and features will be dropped
aswell. see all repositories for the changes done
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
on a side note: the old KUnitConversion implementation was not thread-safe
and was causing plasma-desktop crashes because it was used in the weather
and unit conversion applets, that is no longer the case tho
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
whatever the constraints were attempting to filter and secure is unused now
and I bet in the past - the "Constraints" group is not a thing
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
converting the file to hex can result in way too long file name and
infinite loop in KLockFile::lock()
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
goal is to paint as little as possible with svg images, the fact that there
are 3 kinds of widgets (native, plasma and qml) makes things look
different and as if made of bits and pieces instead of having a uniform
look. with my reimplementation of several widgets and applets it is less
so tho
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
if the image path happened to be the same but the elemente ID not then the
image was not going to change, that is no longer the case
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
while the compression of the scrollbars adjustment has its gain the visual
glitches that resizing on timer causes are not worth it
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
O_CLOEXEC won't do it but saving the locks in /tmp or other directory that
is purged across reboot will remove stale locks
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
the PID check was racy, bonus points for not writing data at all to the
lock and not reading it meaning less disk I/O
oh, yes - by using O_CLOEXEC the lock is stale-safe
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
because the string returned by udev_device_get_action() is valid for as
long as the device is storing it in QByteArray is redundant
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
the service check is racy, KLockFile is as not (assuming the combination of
O_EXCL and O_CREAT is not racy which is not KLockFile's responsibility)
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
so that, for example, the plasma devicenotifier applet does not block
plasma-desktop events when mounting/unmounting
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
