despite the gettimeofday() check for an example, the function is used
unconditionally thus checks for errno header are removed aswell rather then
making them required from configuration checks and error-ing out if not
found.
other cases such as trunc() are used in kexiv2 unconditionally, not in
kdecore.
Signed-off-by: Ivailo Monev <xakepa10@laimg.moc>
netsupp was provided for KDE3 compatibility and the header is not even
installed. the only place where getnameinfo() is used is in KIO kpac DHCP helper
but it does not include the netsupp header.
Signed-off-by: Ivailo Monev <xakepa10@laimg.moc>
TODO from https://github.com/fluxer/katana/issues/7. I've decided to
keep KDE4 and Katie libraries as PUBLIC, all other libraries as PRIVATE
for that and not use INTERFACE at all since that will not require a lot
of changes to all other sub-projects (kde-baseapps, kde-workspace and
kde-extraapps) build systems.
MusicBrainz5 CMake module was also rewritten as the actual library that
should be linked to is libmusicbrainz5cc, libmusicbrainz5 is the C
version of it.
common checks for X11 extensions where move to main config header,
eventually to be moved to separate config-x11 header.
Signed-off-by: Ivailo Monev <xakepa10@laimg.moc>
this allows for less complexity and more abstraction offloaded to
Qt itself which supports dnotify, inotify, kqueue and fsevents.
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
applications search for it in the "locale" resource which points
to the top-level locales directory (usually /usr/share/locale)
and the resource lookups do no do recursive search for it which
means that they would never find it prior to this change
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
it is obviously the wrong approach to report applications crashes
as is noted in the comment in kdebug. since it is only for
KApplications, which provide D-Bus interface, it can be handled
differently with a separate daemon that monitors for wrong
behavior but that is not needed really since DrKonqi does a fine
job at providing feedback when such even occurs
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
there was only one now backend (INI) and changing it on the fly
(as was supported before) can cause serious issues with
applications misbihaving and such
Signed-off-by: Ivailo Monev <xakepa10@gmail.com>
it is bogus with the CMake automoc, requires additional guards in
the source code and some components do not support it at all. Also,
Link Time Optimizations (LTO) seem to be doing exactly what
KDE4_ENABLE_FINAL was supposed to do (basicly) so no point in
trying to support something that modern compilers can already do
for us.
