When using a linear index, all items in a scene are stored in a QList.
While adding new items is a constant operation, removal requires a
traversal through the entire list. This is especially problematic
when the scene contains millions of items and many of them are removed,
which requires a linear search for each item, resulting in a very slow
operation. Moreover, this behavior is actually inconsistent with the
current documentation which states for the linear index:
"Adding, moving and removing items, however, is done in constant time."
With this change, the list is sorted once an item needs to be removed.
The item to be removed is then found using binary search. To reduce
the overhead of sorting the list is not sorted from scratch. First the
newly inserted items are sorted and then the two parts of the list
are merged.
[ChangeLog][QtWidgets][QGraphicsScene] Speed up the removal of items
when using the linear index.
Change-Id: I28708622605d7fb0fac656df1f9b2f2fa3136759
Reviewed-by: Andreas Aardal Hanssen <andreas@hanssen.name>
During the introduction of context adoption support the config choosing got
broken for context creation, at least for hooks that return a customized
format in surfaceFormatFor(). The returned format is the one that needs
to be passed to chooseConfig(), not the original.
Change-Id: Iae203cbbf7b39c462386611dd3744f048116df13
Reviewed-by: Jørgen Lind <jorgen.lind@digia.com>
This reverts commit bf2ec0183c.
That patch changed the way the linear index works in QGraphicsScene
in order to speed up item removal. That patch occasionally rebuilt
the index by recursively exploring all items in the scene, starting
with the top level. Further testing revealved that in some
circumstances, rebuilding the index in this way can be slow, thereby
significantly slowing down the index compared to the unpatched
version.
The original patch only exists in the qt 5.4 branch and has not been
released.
Change-Id: I081dbcdcc86196ef382466c3e800a33eab9a5b79
Reviewed-by: Andreas Aardal Hanssen <andreas@hanssen.name>
This adds a test for a QSslCertificate containing extensions which
are marked as critical.
Change-Id: I314e1f5c9943bcad5d43129a97f9f834882dc6fb
Reviewed-by: Richard J. Moore <rich@kde.org>
Get rid of the rendezvous at shutdown: the android thread does not
need to wait for the GUI thread. Since the GUI thread frequently does
blocking calls to the android thread, this fixes several known and
potential deadlocks.
Task-number: QTBUG-41072
Change-Id: Ia6fa8da026b1727e7352b22f4df4d72b63b8c847
Reviewed-by: BogDan Vatra <bogdan@kde.org>
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
This tightens tests performed on a certificate's extensions by checking
isCritical() and isSupported() for all extensions. It also explicitly
checks the keys when value() returns a QVariantMap.
Change-Id: If51c55be25bbcd09cc3a6712ddfea2bf9a01360f
Reviewed-by: Richard J. Moore <rich@kde.org>
There are valid uses cases how to use Qt without installing
any fonts (e.g. by using an application font), so let's
make the warning non fatal.
Task-number: QTBUG-29192
Change-Id: I5684331b687bef4b8a54be1f68303c80aa8d4372
Reviewed-by: Paul Olav Tvete <paul.tvete@digia.com>
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
Assignment of a value was two times successively for same variable while
the variable itself is not used between these assignments.
Change-Id: I3c457e3af0505d32a64f6c8576b458cd15a951e5
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
The practical uses of a POD QElapsedTimer are not really that clear, and the
number of misuses of this API are quite high. Default the state to invalid to
prevent against mistakes.
[ChangeLog][QtCore][QElapsedTimer] Is no longer a POD.
Change-Id: I267292acf2bfca7404e3e449dd04410441d7ce26
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
udpSocket() can only be called after the socket descriptor is set.
Change-Id: If651ff58507cd66c98de4b699a80149913d9ffcf
Reviewed-by: Andrew Knight <andrew.knight@digia.com>
Move the line that adds the external page definions into its own
separate include file.
This prevents the files from being included in every Qt
documentation module by default, needlessly duplicating the
information. Currently only the QtDoc module (and modules that
depend on QtDoc) need it.
Task-number: QTBUG-41003
Change-Id: Ie2ddd3a645ea731787daacfffb3068cb36c79c9a
Reviewed-by: Martin Smith <martin.smith@digia.com>
Reviewed-by: Venugopal Shivashankar <venugopal.shivashankar@digia.com>
This adds the infrastructure for reading and writing encrypted private keys
when using non-OpenSSL backends. Each platform must provide its cryptographic
encrypt / decrypt functions.
As WinRT already uses the common parser, this commit includes an
implementation for that platform.
Done-with: Andrew Knight <andrew.knight@digia.com>
Task-number: QTBUG-40688
Change-Id: I0d153425ce63601ff03b784a111e13962061025f
Reviewed-by: Richard J. Moore <rich@kde.org>
Do not try to resize or destroy invalid surfaces.
This caused update problems with all GL apps after suspend,
since we would forget the dummy view that we always keep
around so we get proper transitions on shutdown.
Also make sure that we don't mess this up even if we try to destroy a
non-existing surface. This would have fixed the bug by itself, but
then we would still be stuck with the annoying warning message.
Task-number: QTBUG-41093
Change-Id: I83299e93eb9ac5357b98ca47014789b56c91b35a
Reviewed-by: Christian Stromme <christian.stromme@digia.com>
The default behavior of ANGLE is to use D3D11 before falling back to D3D9.
This change improves flexibility the platform plugin to explicitly create
a D3D11, D3D9, or D3D11 software (WARP) context by setting the
QT_ANGLE_PLATFORM to "d3d11", "d3d9", or "warp", respectively.
Task-number: QTBUG-41031
Change-Id: Ie1d399c1cb0e360e5b3a6d9f2a4b28745d86cc71
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Laszlo Agocs <laszlo.agocs@digia.com>
The color table is passed by value (good for C++11), so when
the argument is assigned to the member variable, that's a good
spot for a move assignment.
However, the argument is also declared const.
The standard says that top-level const is ignored, but some
compilers (I know about SunCC) think differently, so we cannot
remove it.
Instead, we do a const_cast. It is well-defined: Even though
apparently the argument was declared as const, the standard
says the const is not there, and no sane compiler would put
the argument copy into read-only memory.
Add a reminder to remove the top-level const from the
signature come Qt 6.
Change-Id: Iac18846ba669de0a30da620685ad1438c267e193
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
The code doesn't touch any member variables, so it doesn't need mutex
protection. Reducing the time spent under the mutex allows a higher
speedup (Amdahl's Law), so do it.
Also made the mutex locker const to indicate it is never unlock()ed
again.
Change-Id: Ic50b827c0e34d39cbddc7ec83675b568a9c33f6d
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@digia.com>
Fixes missing space before `]` in test condition introduced in 8d57725338.
Change-Id: I741c291677f32056a0a0bec12cb4d9fd293a2021
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
We pull out the magic UIViewAnimationCurve of the keyboard animation
when the keyboard is about to show, but we need to defer the shifting
of the value 16 bits, as that turns it into a UIViewAnimationOptions,
which we can't store in a UIViewAnimationCurve member.
Change-Id: Id35dae1ec487951df749dfffb6118b572c28b103
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@digia.com>
Extract a struct containing the function pointers and derive
QWindowsXPStylePrivate from it, so that both styles can use them.
Remove duplicated variables.
Task-number: QTBUG-38858
Change-Id: I7c2c665b5930c56ffdf33c5185720f71517d146c
Reviewed-by: Alessandro Portale <alessandro.portale@digia.com>
This change adds the ability to decode ASN.1 INTEGER fields,
provided they represent a positive number of less than 64-bit.
This is needed for PKCS#12 decoding.
Change-Id: Iafb76f22383278d6773b9e879a8f3ef43c8d2c8f
Reviewed-by: Oliver Wolff <oliver.wolff@digia.com>
Otherwise bad things happen if the application leaves a buffer bound.
[ChangeLog] The eglfs mouse cursor properly resets the array buffer
to prevent rendering issues. This, just like with vertex attributes and
textures, requires applications to be aware of such state changes in
swapBuffers().
Change-Id: I8b383cc867d8d0d0572773eacfa650e7b33b9680
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
It's easy to do the work on the side and then commit. This is strongly
exception safe, but in Qt, we don't care. But transactional code, when
this easy to achieve, is also clearer.
Change-Id: I30f1badec7745d62a09af4eede234cb312b373aa
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
When using WA_TranslucentBackground/FramelessWindowHint, the backing
store must paint to an offscreen texture instead of the swap chain in
order to achieve the desired results. This texture is then presented to
the screen using UpdateLayeredWindowIndirect().
As the swap chain is not needed in this mode, its construction is skipped
if indirect rendering is active.
Furthermore, the layering options were updated to fix an issue with
transparent layers overpainting the background. The layer options were
switched to D2D1_LAYER_OPTIONS1_NONE, which appears to work for both
translucent and non-translucent rendering modes.
Task-number: QTBUG-40601
Change-Id: I656f7cdfb424d1eda6f82c2c69500e78d8c1726a
Reviewed-by: Louai Al-Khanji <louai.al-khanji@digia.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
The scale factor was only being applied to touch coordinates.
Change-Id: I7fc2793b1514c73986a574a95478306c1eb54c5e
Reviewed-by: Oliver Wolff <oliver.wolff@digia.com>
Allow the D3D11 renderer to build with the June 2010 DirectX SDK.
Change-Id: I2343acedab16845d6a0d4a53cf3145f583efc4a7
Reviewed-by: Andreas Holzammer <andreas.holzammer@kdab.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
These libraries are dynamically loaded on desktop Windows.
Change-Id: I3a0d17a48a3bd4930690d20d387df0d92906662d
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Platform plugins have a tendency to make the newly created native
context current with a temporary surface. This is usually needed to
query some information related to the new context. Afterwards most of
them just reset to having nothing current.
This has two issues:
It unexpectedly changes the current context/surface. A call into
QOpenGLContext::create() does not imply that the current context will
get changed. This is the minor issue and we could probably live with
it (at least if it had been documented).
However, the real issue is that QOpenGLContext::currentContext() will
become inconsistent: it will still report whatever was current before
the create() even though on the EGL/WGL/GLX level that's not the case
anymore.
To prevent all this confusion the platform plugins can easily be
changed to restore whatever context/surface was current before they
altered it.
Change-Id: I6a5b4597c86571327524ddb13e0d02538593cc7b
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
Some fonts appear to include zero-width space in the CMAP table, but
not include an actual definitions of the glyph they point to. The
missing glyph causes a warning, but isn't handled making the character
end up being giving the same metrics as whatever character it came after.
This patch adds explicit handling of missing glyphs, and also caches
their missing state when caching is enabled.
Task-number: QTBUG-40912
Change-Id: I06fba9c01df59548e750e36babfdd5a6bafd6bd0
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Pierre Rossi <pierre.rossi@gmail.com>
That commit exposed part of qfunctions_winrt.h for non-WinRT use
(Windows 8) without wrapping with Q_OS_WIN. That meant the headercheck
pass failed to compile when outside of Windows.
Change-Id: Ie731cce21e5102f5e5879b147b7738a7d0a91ecd
Reviewed-by: Andrew Knight <andrew.knight@digia.com>
QSslKey currently has methods which supposedly allow decoding and
encoding private keys as DER protected by a passphrase. This is
broken by design as explained in QTBUG-41038, as storing the encrypted
DER data alone makes no sense: such a file lacks the necessary
information about the encryption algorithm and initialization vector.
This change:
- explicitly stops using the passphrase when decoding DER in the
constructor. The behavior is unchanged, it is not possible to
read the encrypted DER alone.
- refuses to honor the passphrase to DER encode a private key. The toDer
method now outputs an empty QByteArray instead of garbage.
Task-number: QTBUG-41038
Change-Id: I4281050cf1104f12d154db201a173633bfe22bd9
Reviewed-by: Richard J. Moore <rich@kde.org>
This adds a test for 3DES encrypted keys in addition to the
current DES encrypted keys.
Change-Id: I229e3ef710e9ee23efa2a3275b89d958491de4a2
Reviewed-by: Richard J. Moore <rich@kde.org>
