| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
| |
* removed the vr41xx, yopy and sl5000 drivers (old PDAs)
* removed the bus mouse driver (ISA days should be over)
* renamed the LinuxIS mouse driver to LinuxInput (consistency with the key driver)
* unified the LinuxInput mouse and key driver I/O handling
Reviewed-By: Paul
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| | |
Using qglobal.h and checking the compiler version with the
preprocessor has the side-effect that moc won't generate proper code
since it doesn't know about the compiler version.
Enable both modules under Sun CC 5.9 and IBM xlC 7.0.
|
| | |
|
| |
| |
| |
| |
| |
| |
| | |
'y' rather then 'yes'
Merge-request: 945
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
|
|/
|
|
| |
Reviewed-by: Lincoln Ramsay
|
|
|
|
|
|
|
| |
It was reported to be auto-detected, but wasn't. Apparently, AIX 6's
X11 doesn't have this.
Reviewed-By: Denis Dzyubenko <denis.dzyubenko@nokia.com>
|
|
|
|
|
|
|
|
|
| |
-developer-build, part 1.
Adds QT_CONFIG+=private_tests to qconfig.pri when Qt is configured
with -developer-build.
Reviewed-by: Michael Goddard
|
|
|
|
|
|
| |
.../configure: line 4462: [: missing `]'
Reviewed-by: Jason McDonald
|
|\ |
|
| |
| |
| |
| |
| |
| |
| | |
ShivaVG runs on top of OpenGL. Include the QMAKE_*_OPENGL flags
on the link line if it is needed by the OpenVG engine.
Reviewed-by: trustme
|
| |
| |
| |
| |
| |
| |
| |
| | |
We have said in the past that 10.3 has been dropped for 4.6, but we
actually have to do the work and make sure that it happens. This is the
first step.
Reviewed by: Morten Sørvig
|
| |
| |
| |
| |
| | |
This change also moves the EGL support classes from QtOpenGL to QtGui
so they can be shared between OpenGL and OpenVG.
|
|/
|
|
| |
Reviewed-by: Bill King
|
|\
| |
| |
| |
| |
| |
| | |
Conflicts:
configure.exe
tests/auto/qpainter/tst_qpainter.cpp
translations/qt_ru.ts
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
it simply breaks too much (embedded builds in particular).
will come up with something better in master.
note however, that the .qm files do NOT return to the repository. the
release manager needs to run "make qm" before packaging, just like
Qt-from-git users who want translations do (i.e., almost nobody).
Reviewed-by: Jason McDonald
|
|\ \
| |/ |
|
| |
| |
| |
| | |
Reviewed-by: Trust Me
|
| |
| |
| |
| | |
Reviewed-by: Trust Me
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
so "qmake -r && make" worked fine ... but configure does not simply run
"qmake -r".
also, now we have to build lrelease even when we don't build the other
tools.
Reviewed-by: Jason McDonald
Reviewed-by: mariusSO
|
| |
| |
| |
| | |
Reviewed-by: TrustMe
|
| |
| |
| |
| |
| |
| |
| | |
This gets rid of a few warnings when shadow-building Qt, since the
symlinks are no longer created in the source tree.
Reviewed-by: Paul Olav Tvete <paul.tvete@nokia.com>
|
| |
| |
| |
| |
| |
| |
| |
| | |
This makes sure that we only detect and use libraries that are actually
in the SDK.
Task-number: 109757
Reviewed-by: Richard Moe Gustavsen
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Some older versions of the Bourne shell (/bin/sh) don't
support the ! operator for negating the exit code of a pipeline.
In particular, this is true on certain (all?) Solaris 9 platforms.
The patch replaces all occurrences of the ! operator with equivalent
(but slightly less elegant) expressions that test the $? variable instead.
Reviewed-by: msorvig
|
| |
| |
| |
| |
| |
| | |
If unixodbc detection fails, try and link/detect iodbc libraries and link against them.
Reviewed-by: Lincoln Ramsay
|
|\ \
| |/
| |
| |
| | |
Conflicts:
tests/auto/qtreeview/tst_qtreeview.cpp
|
| |
| |
| |
| |
| |
| |
| |
| | |
Instead of the multiple character-string replacements, just check for
the discrete -arch values that we want to find. This makes the code
clearer and should reduce the chance of subtle errors.
Reviewed-by: Jason McDonald
|
|\ \
| |/ |
|
| |
| |
| |
| | |
Reviewed-by:cduclos
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The previous "fix" caused the effect of the line before the change to be
discarded, which breaks the mac binary package builds. When reviewing
changes, please read the lines of context given in the patch - they are
included for a reason.
Acked-by: Lincoln Ramsay
Acked-by: Rohan McGovern
|
|\ \
| |/ |
|
| |
| |
| |
| |
| |
| |
| |
| | |
The reason is the tha config.test ssl failed building. And the reason for
that was a bad makefile flag (i386_64).
Task-number: 253887
Reviewed-by: Trenton Schulz
|
|\ \
| |/
| |
| |
| |
| | |
Conflicts:
src/gui/painting/qbackingstore.cpp
src/gui/painting/qwindowsurface_raster.cpp
|
| |
| |
| |
| |
| |
| |
| | |
STRING="${STRING/x86/i386}" actually replaces whole words only,
so we don't need to fixup x86_64 afterwards.
Reviewed-by: nrc
|
|\ \
| |/
| |
| |
| |
| | |
Conflicts:
src/gui/kernel/qcocoaview_mac_p.h
src/gui/widgets/qmainwindow.cpp
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Store whether qt was built with qt3support enabled in .qmake.cache. We
need this info to decide if QT3_SUPPORT should be defined when we
include qimage_p.h. Otherwise we get data corruption issues since the
size of struct in Qt is larger than the one we include in libqdirectfb.
Reviewed-by: Donald <qt-info@nokia.com>
|
| |
| |
| |
| |
| |
| |
| |
| | |
Reported downstream to Debian:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=525570
Reviewed-by: Alexis
Reviewed-by: Thiago
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
architechtures.
For example,
./configure -arch x86_64 -arch ppc requires D-Bus compiled for x86_64 and ppc
in order to enable it.
Task-number: 249881
Reviewed-by: nrc
|
|\ \
| |/
| |
| |
| | |
Conflicts:
src/gui/itemviews/qabstractitemview.cpp
|
| | |
|
|\ \
| |/
| |
| |
| | |
Conflicts:
src/gui/itemviews/qheaderview_p.h
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The nightly public snapshots contain bin/syncqt but do not
contain the include/ directory. Hence configure needs to
detect this case and run syncqt to create the include tree
so that configure actually passes
Signed-off-by: Thiago Macieira <thiago.macieira@nokia.com>
|
|\ \
| |/
| |
| |
| | |
Conflicts:
tests/auto/qpainterpath/tst_qpainterpath.cpp
|
| |
| |
| |
| |
| |
| |
| |
| | |
The previous logic caused a license check when the eval
license file was missing.
Reviewed-by: Jason McDonald
Task-number: 251339
|
| |
| |
| |
| |
| |
| |
| |
| | |
The eclipse integration is namespaced, but it still pulled in system plugins like the oxygen style through QT_BUILD_KEY_COMPAT check in QLibraryPrivate::isPlugin()
Task: 250185
Reviewed-by: dt <qtc-committer@nokia.com>
BT: yes
|
| |
| |
| |
| |
| |
| |
| |
| | |
The eclipse integration is namespaced, but it still pulled in system plugins like the oxygen style through QT_BUILD_KEY_COMPAT check in QLibraryPrivate::isPlugin()
Task: 250185
Reviewed-by: dt <qtc-committer@nokia.com>
BT: yes
|
|\ \
| |/
| |
| |
| | |
Conflicts:
src/gui/graphicsview/qgraphicsitem.cpp
|
| |
| |
| |
| |
| |
| |
| | |
The previous version would get confused because the embedded package
also contains _x11 files.
Reviewed-by: Thiago
|
| |
| |
| |
| |
| |
| | |
GCC uses i386, but configure has always used x86, which can lead to confusion.
Reviewed-by: nrc
|
| |
| |
| |
| | |
Reviewed-by: TrustMe
|