The normalize function don't do well it's job.
Given a path like that, the regexp "/[^/]+/.." will match 3 times
/usr/lib/gcc/x86_64-unknown-linux-gnu/4.8.2/../../../../include/c++/4.8.2
[ ][ ][ ]
The second match is wrong as it will remove /../..
Use sub instead of gsub which only remove one match at the time.
Change-Id: I0657bc603e521c9e53b9f50d2481dce184b64bad
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
... except where they actually contain variable expansions.
unescaped backslashes in double quoted strings aren't nice, and
apparently actually break with old solaris shells.
Task-number: QTBUG-14167
Change-Id: I703694b6ac7ab71d9293c049d77212e20dd5bcb0
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Passing -no-xcb disables xcb and related configuration tests
completely so for example QT_NO_XRENDER is not defined. This results
in linker errors in glxconvenience with -no-xcb since the XRender
calls are compiled in without linking to the library.
The XRender, XInput and such tests are not strictly xcb related and
may be used from elsewhere, for example glxconvenience which has
nothing to do with xcb. Therefore these tests are moved in configure
so that they are run even when xcb is disabled.
Task-number: QTBUG-35644
Change-Id: I77871612ea5f6711ecafd8ca53aac6c516c19e2e
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Needed for using alternate stdlib implementation like libc++ with clang.
Change-Id: I1782f62f5e2ea95e6cff8a1ed646362c0a31645c
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
This is a left-over from the cleanup of MMX/3dNow!/SSE support, which we
no longer have in Qt.
Change-Id: I48388710a499bddb518ae3c2b8a4ad989482f58c
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
If SSE2 isn't supported, then SSE3 can't be either. Onwards and upwards
for SSSE3, SSE4.1, SSE4.2 and AVX. The test for AVX2 was already there.
Task-number: QTBUG-24773
Change-Id: I005258db52d8abcd407a99b8ebcc23cdea8e3d9f
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
d34cae51 introduced a new configuration parameter, xcb-xlib, but did not
remember to set the internal variable after the config test was run,
resulting in a potentially incorrect reporting of the option.
Change-Id: I979589d9aad2eb5e0fac069dfcd5b329a51ae059
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
the change is wrong for multiple reasons:
- it is possible to build qt (the offscreen plugin) with xlib but
without xcb, which was impossible after the change
- the check was built in a way that if xcb is auto-detected but xlib is
missing, configure would abort (instead of disabling xcb), which is
unreasonable (cf. linked task)
- it should be possible to build the xcb plugin without xlib - it's
testing for xcb-xlib and should have the correct #ifdefs. if these are
broken, they should be fixed. if this is unrealistic, the build system
should be adjusted to this fact ... differently.
This reverts commit 683451e7c7.
Task-number: QTBUG-34382
Change-Id: I1c5811e2fb1f09b084b6aeb7350e5f376e1b9b82
Reviewed-by: Nicolás Alvarez <nicolas.alvarez@gmail.com>
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
Testing feature macros is not enough for uclibc. Fixes build of the built-in
sqlite3 with uclibc <= 0.9.33.2. Later versions will have posix_fallocate().
Change-Id: I918a52777ac63624635802221effc6b86fa2269c
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Reviewed-by: Mark Brand <mabrand@mabrand.nl>
We require glx version 1.3 which is where fbconfig is first defined.
Also make use of the configure test and report the glx status.
GLX support should always take precedence when compiling the
xcb backend
Change-Id: Ie46834210bf5cd2ac4006ff08379e0d3434ffa2b
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
This adds a new configuration parameter, xcb-xlib, which allows
overriding the configure test for that existing configuration option.
The use of xcb-xlib in the xcb platform plugin becomes the preferred path
for non-OpenGL ES 2 builds, while the EGL codepath is used otherwise.
This has the advantage that EGL can be used with Desktop OpenGL if Qt is
configured with -no-xcb-xlib.
Change-Id: I5018e31fe0399b94f020c671eff9414d00431c44
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
desktop opengl and egl are not incompatible with eachother, so there
is no need to disable egl when on desktop opengl
Task-number: QTBUG-34949
Change-Id: I757c38674a480910d1cb23853c255eb993e107ff
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Reviewed-by: Andrew Knight <andrew.knight@digia.com>
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
The QtGraphicalEffects module is a QtQuick2 only module so it is clear
why it was left out before iOS had support for QtQuick2, but now it
should be built by default. This commit removes QtGraphicalEffects from
the iOS module "skip list".
Change-Id: Ie0890f6784ad6e276e80bf7a11195c141ca23332
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
The QtQuickControls is supported for 5.2 though it does not provide
the system native "look and feel". This patch removes the module from
the iOS module "skip list".
Task-number: QTBUG-35033
Change-Id: Icad479f2e4651c02e569e8a86cfe5ac912d8ed10
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
QtMultimedia is supported in Qt 5.2 and should be built by default, so
it has been removed from the module "skip list"
Task-number: QTBUG-35039
Change-Id: Id43f997251bf392b7f6c0166ee06b7364ce30b8e
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
it's beyond me how it landed in the middle of the sql driver detection.
Change-Id: I7d1302e3440b8828a291e3009e9b130a2fdf165f
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
Plain QNX 6.5.0 does not have a libpps, the new QNX
has a libpps and BlackBerry has it as well. So we need
a configure check to not open another mkspec for this
platform. This fixes the plain QNX 6.5.0 build.
Change-Id: Id4b3876f2385bcb5f3df426945532e7e26133f24
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>
introduce XPLATFORM_MAC which is properly initialized from the spec
instead of from the host.
use that and BUILD_ON_MAC where appropriate.
minor bug: during command line arg validation we use BUILD_ON_MAC even
when XPLATFORM_MAC would be in order, because the latter is not
available at that point yet. the solution would be delaying the
validation, but that doesn't seem worth the effort now.
Task-number: QTBUG-33896
Change-Id: I63f361212961bfdd82efde2ca71a1f48904a85fb
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
... which is right after performing the platform detection, as proper
license checking needs to know the target platform.
we can do that now, as we moved all more expensive or side effect laden
processing to a later point.
as a side effect, we also get rid of the weird early cmdline parsing.
Change-Id: I0fda0a15a1ea2082603f7097e89aa422853f30fa
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
is the only sane thing, and now that there are no dependencies on
anything else any more (in particular anything slow or side effect
laden), we can actually do it.
this removes some nasty OPT_HELP checks spread throughout the code, thus
further reducing data dependencies.
Change-Id: Ib0a00e1514e2aa25ec3b527ba9f5719e3214640b
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
the help would echo the defaults adjusted by command line overrides and
environment variables for some options. this was entirely pointless.
it also printed the result of the os detection, which was mostly
useless.
but the primary reason for doing this is reducing data dependencies,
so the code can be moved up.
not touching the windows configure, as it's a lot harder and currently
that behavior isn't in the way of what i need to do.
Change-Id: Ide62119091c8494fb3bea2b607f140268f057919
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
it's counterproductive to do it earlier: it's passed down to configure
tests, which then log less info than they could.
but primarily, this serves the purpose of minimizing the amount of code
executed before platform detection.
Change-Id: Iff19b8555de19d048ea6d9341af965871b314c54
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
minimizing the amount of code executed before platform detection.
Change-Id: Ib2c0d97ce5040ced8c4c41d74f428fe7d0f75664
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
it's entirely pointless to flood the user with information and force him
to scroll back when he most likely just made a typo.
apart from that, this reduces the data dependencies, thus easing further
refactoring.
Change-Id: I7b24274d453de54a4f02481a66d77e27d4ab0657
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
moving dead code out of the way of further refactoring.
Change-Id: If558406cdf13d61478634dd7eff644dc67b0e53d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
have this option independent code out of the way before starting option
processing.
Change-Id: I5a08caeb25689b155c256ef82505c000112f5039
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This way, a Qt compiled with qreal=float and one linked
with qreal=double can not be linked by a single downstream. That is
diagnosed at cmake-time.
Change-Id: I9183dbcfef181fadea5321d3154948e8258e4a2a
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
On modern ARM CPUs there is no speed difference between
float and double anymore, so let's rather use double for
qreal to avoid rounding and precision issues. Like this
we also get much better compatibility with our desktop
OSes.
This is not binary compatible on ARM, but the old behavior
can be restored by passing -qreal float to configure.
Change-Id: I2a4b61e19a3dfa6b0bd76734cecf2634c97207fc
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
this makes it possible to properly exclude entire subprojects based on the
availability of features, rather than stuffing every single source file
with #ifdefs.
the defines are aggregated from the -qconfig <profile>,
-no-feature-<foo> and some other configure flags.
usage:
load(qfeatures)
!contains(QT_DISABLED_FEATURES, textarea): SUBDIRS += textstuff
Task-number: QTBUG-28102
Change-Id: I83400632d64312fa4b907e1318dddfe27c432387
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
Reviewed-by: Peter Hartmann <phartmann@blackberry.com>
Reviewed-by: Tasuku Suzuki <stasuku@gmail.com>
the path filtering functions assume that DEFAULT_{INC,LIB}DIRS are
newline separated lists of unquoted strings, which 8fbf959be broke.
Task-number: QTBUG-33714
Change-Id: Ie07909963ac5155a8ac79ca9254f34069925e001
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Conceptually a Qt for iOS SDK or source build should support building
for both simulator and device, based on the same qmake binary and Qt
libraries. Qt Creator or Xcode should then be able to use the same Qt
version while still building for a single target at a time. This
applies to user libraries as well, which shouldn't require switching
to a different Qt when changing target platform from simulator to
device.
We achieve this by using Qt's exclusive_build feature, where we build
for the two targets in parallel, and then teach Xcode how to choose
the right library dynamically at build time.
Change-Id: I06d60e120d986085fb8686ced98f22f7047c4f23
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
We used to compute the default exclusive build directory, eg 'debug', at
configure time, and then set OBJECTS_DIR, MOC_DIR, etc to include this
hard-coded default exclusive build directory. We then had to run a post-
process step where we replaced the 'debug' part with the current actual
exclusive build pass, eg 'release', resulting in long-standing bugs such
as QTBUG-491 where we end up replacing parts of the build output dirs
that were not part of the original exclusive build directory.
We now set the OBJECTS_DIR, MOC_DIR, etc defaults in configure like
before, but they do not include any exclusive-build information. The
exclusive build directory is handled as a separate step in default_post
where we adjust all entries in QMAKE_DIR_REPLACE to be exclusive
directories.
For backwards compatibility the new exclusive build behavior is only
enabled for variables named by QMAKE_DIR_REPLACE_SANE, which for Qt
itself applies globally to everything but DESTDIR, and for libs and
tools also applies to DESTDIR. The reason for leaving out DESTDIR in
the general case is because many tests and examples assume the old
behavior for DESTDIR. A side effect of including all the other
variables for Qt libs and tools is that the PCH output dir will be
uniformly set, which has been an issue on Windows in the past.
The addExclusiveBuilds function now takes two or more arguments,
each argument being the key for an exclusive build, which can be
customized eg. using $$key.{name,target,dir_affix}. Passing more
than two arguments results in three/four/etc-way exclusive builds,
eg debug/release/profile. Exclusive builds can also be combined, eg
static/shared + debug/release by making two calls to the function.
We also handle individual targets of combined exclusive builds,
eg static/shared + debug/release, meaning it is possible to run
'make debug' to build both static-debug and shared-debug.
Task-number: QTBUG-491
Change-Id: I02841dbbd065ac07d413dfb45cfcfe4c013674ac
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Status for LIBUDEV printed again in the configure output under 'udev ....'
Change-Id: I3e0beca4c6de315dc2dfbdc24d83f4d9cd2b8856
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Nothing in qtbase uses gstreamer, so this is just a red herring.
Change-Id: I93fb20a70928d84fed8f33ca4c5df38779928f1a
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
apparently some awks need the regexp delimiters to be escaped even
within a character list.
Change-Id: I74a77022e134cafa269960c5c88a9a88589f1b95
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
