| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Using QTextCursor::Left and QTextCursor::Right moves the cursor visually to right or
left, which does not work with Right to left text, where moving left means advancing to
the next character, and vice versa.
By using QTextCursor::PreviousCharacter and QTextCursor::NextCharacter we leave it up to
the engine for deciding where to move depending on the context.
Task-number: QTBUG-50619
Change-Id: Ibf0a886f3ff6bf9ca2c953aeff8c235612bbc01d
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
|
|
|
|
|
|
|
|
| |
Fixed breaks caused by MSVC 2017 15.8 introducing the
__cpp_enumerator_attributes define.
Change-Id: I78144f8f49a7352e574dee379155bd47d8d6d896
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
| |
If we ever need to add QCborValue to the bootstrap library, it's
unlikely that we'll need this part. And by splitting it, I can make the
code handle more cases, that hadn't been properly handled before.
Change-Id: I2f630efbbce54f14bfa9fffd154160c0ad893695
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
|
|
|
|
|
|
|
| |
The broke compilation with clang on linux.
Change-Id: I4731017d8a89c11821fbdd98128e27aafc857f97
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|
|
|
|
|
|
|
|
|
|
| |
There is no such file, if one of the "directory" components of its
path is not, in fact, a directory. Added a test for non-existent file
(specified to give empty canonical file path) as well as a test for a
file in a sub-directory of a known file. The former incidentally
tests for QTBUG-29402, fixed long ago.
Change-Id: I60b80acc0f99f0a88cdb1c4d191af7384f3a31c5
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
| |
Makes clear that this is what it is; and ensures we'll get compiler
warnings if someone adds a new entry to the FileName enum without code
to handle it here.
Change-Id: I36e383066728cefcc75e0a760e36222cebd1dff0
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
On Unix, we wouldn't even *try* to truncate if the file was open for
appending. The combination may be an eccentric choice but - at least
when it's combined with reading - I can imagine use-cases for it; and
we should (at least try to) deliver what we're asked for, even if we
can't think why anyone would want it. So actually enable truncation
when asked to.
Amended some tests to check this works and corrected the QIODevice
documentation of mode flags (which misdescribed the special case that
implies Truncate). Removed special-case code, to apply truncate when
writing but not reading, since it's been made redundant by the
pre-processing of mode done in QFSFileEngine::processOpenModeFlags().
[ChangeLog][QtCore][QFile] When opening a file, if Truncate is asked
for, or implied by other flags, it shall be attempted, regardless of
what other options are selected. We previously did this on Windows;
now we do so also on Unix (even when appending).
Task-number: QTBUG-13470
Change-Id: I1e08d02cfbae102725fccbbc3aab5c7bf8830687
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
| |
Change-Id: I6d0bf78d02d166307f864f1f83a3b600ef6a9b0b
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
|
|
|
|
|
| |
Change-Id: I8fc1ba7aac1c8ac86a8cb5d6f864180e0721926f
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
|
|
|
|
|
|
|
|
|
|
| |
It was supposed to be used in Qt Quick Controls 1, for the desktop
style, but the followup patches in QQC1 never landed, and QQC1 is
now deprecated.
Change-Id: Iceefd523fc02a9e48b986dc33bb13a41804dd199
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
|
|
|
|
|
|
|
|
|
| |
As now the feature 'dtls' depends on the feature 'openssl' - ifdefs
are redundant, this code is always 'openssl-only'.
Change-Id: I6a7fe9e3a00ae05656af1626e7de74e813df5d32
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
|
|
|
|
|
|
|
|
|
| |
__has_include(<compare>) is not the correct way to detect this feature,
since that's a library header and may be provided by an implementation
(libc++) before the compiler supports the syntax.
Change-Id: I80aae0d068974d83b6c0fffd1544c8e558e2446b
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
For keys, QSettings escapes all characters outside of [-a-zA-Z0-9_.]
by using percent encoding, and changes '/' to '\'. That is,
settings.setValue("qt.*", true)
will be written to an .ini file as
qt.%2A=true
This means that QSettings can not be used to write general-purpose
qtlogging.ini files. Fix this by applying the reverse transformation
method from QSettings when reading in the .ini file.
[ChangeLog][Logging] Qt will now accept qtlogging.ini files
written by QSettings.
Task-number: QTBUG-69548
Change-Id: I55b7a8b433291268dc6855901f72b1c04f8ee6d3
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
| |
Allow to use binary search instead of iterate over array
Change-Id: Ibcce1a2296e85f51a61fbb5e64e23e5a27464959
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
|
|
|
|
|
|
|
|
|
| |
Unfortunately, this enum is not 'binary', it also has 'UnencryptedMode'
and as a result we can end up with server-side method/context.
Change-Id: If2da4c1b9f7e9ff916d933c9517c00a632aea324
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
|
|
|
|
|
|
| |
Task-number: QTBUG-25550
Change-Id: I37ec02b655abe2779aa11945e20550ce00e43723
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
| |
The emulation was not working since the vertices are clamped before
getting to the fragment shader. So instead just resize the brush if
not supported.
Change-Id: I856e47890cd3021874b77d869a6ff7162cadde10
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
|
|
|
|
|
|
|
|
| |
Turns out on OpenGL ES, only the GL_RGB10_A2 form is allowed as a
render buffer storage format.
Change-Id: I42915b61835167ae457aae91da7e75065dd3eb21
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
|
|
|
|
|
|
|
|
|
|
|
| |
This way the logic can be reused elsewhere.
At the same time a standard OpenGL ES/3 way of handling BGRA is added,
so we don't depend on extensions, and handling of NPOT and max size
which QSGTexture will need.
Change-Id: I475bc7127f44be3964fdb482c9e86a20db1fbca5
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The ascending variable was a leftover from an old implementation. Now
it's not needed anymore. It's safe to remove it, and the condition can
be removed as well.
Amends commit 699e8fe3a6724dd8288dd6a55b25e070eb198c9d
Coverity-Id: 11004
Change-Id: I003b7016e9bd3bd523fb70ed4de84211ec840a78
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Samuel Gaist <samuel.gaist@idiap.ch>
|
|
|
|
|
|
| |
Change-Id: I719856168a254c90b0bb0943faee5ab0be75b0bd
Reviewed-by: Martin Smith <martin.smith@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
|
|
|
|
|
|
|
|
| |
More Qt-style and more natural, also, shorter names.
Change-Id: I97bd68a8614126d518a3853027661435dc4e080d
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
After a handshake was completed, TLS socket is in 'connectionEncrypted' state.
So on a read notification, in 'transmit', we call 'SSLRead' to read supposedly
encrypted application data or TLS internal messages. In case SSLRead finds either
ClientHello or HelloRequest from a server, it attempts in a rather sneaky manner
to renegotiate. And as it happens here and there with SecureTransport, SSLRead
fails and the work is only half-done, since we have kSSLSessionOptionBreakOnServerAuth
and kSSLSessionOptionBreakOnCertRequested options set to 'true'. We end up with
completely unexpected errors like errSSLClientCertRequested or errSSLPeerAuthCompleted
(yes, this is so normal and totally expected for 'SSLRead' function to verify
certificates and WRITE messages, no need to document this at all!).
If SecureTransport is sneaky, so can be us:
- in a read callback SecureTransport is probing the type of record
and we can notice a sudden session state change - it goes from
kSSLConnected (which is set upon handshake completion) to
kSSLHandshake (which means a (re)handshake is ongoing);
- if this is the case - we lie to SecureTransport about the amount
of data available (0 bytes), set 'renegotiating' to 'true', return
errSSLWouldBlock;
- in 'transmit', if SSLRead returns errSSLWouldBlock and 'renegotiating'
was set, we call 'startHandshake' until isHandshakeComplete() == true
or some error encountered.
[ChangeLog][QtNetwork][QSslSocket] Implement renegotiation for SecureTransport backend
Task-number: QTBUG-69420
Change-Id: Iaab1336aa3abf3f6ac94b358f3142d2738a18ee9
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
|
|
|
|
|
|
|
|
|
| |
Keep the original class name around for a little longer so we can
generate the correct scoped enum in the moc output.
Task-number: QTBUG-47652
Change-Id: Ib5934316fa786cc475335b03c86b8ec2dc239055
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When we have the named keys and not just integer values, we can output
something unambiously that closer match how the enums should be used.
Output of enums without proper metadata is left unchanged
Before:
QSurfaceFormat::ColorSpace(DefaultColorSpace)
QPainter::CompositionMode(3)
After:
QSurfaceFormat::DefaultColorSpace
QPainter::CompositionMode(3)
Change-Id: I537e879ba8b5c555b2aae9ba831facc88d430443
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
|
|
|
|
|
|
|
|
|
| |
This patch renames rather awkward 'remote' into more conventional
'peer' (similar to what we have in QAbstractSocket).
Change-Id: Ifc45e538b8adf9cc076bd7aee693277829fd94dc
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
| |
When an element is not visible then it should not be shown in the element
tree at all when using tools like Inspect. Also, the IsOffscreen property
was hardcoded to false instead of reflecting the actual object offscreen
state.
Task-number: QTBUG-69537
Change-Id: Ic8f55486685837cf5e21b3499085bb669c1dc6c8
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The --list option outputs only the file system paths, but does not tell
us where those files end up in the resource file system. This is
unfortunate because the resource paths are possibly the most interesting
thing to be expected from a listing of resource files. Add an option
that outputs a full mapping.
This is also useful for tools that need to resolve files for resource
paths from outside the application using the resources.
[ChangeLog][Tools][rcc] Added a --list-mapping option which shows a
mapping of resource paths to file system paths.
Change-Id: I0cb57fe091ef78b4a9eca8be2fc53278ae50cc11
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: hjk <hjk@qt.io>
|
|
|
|
|
|
|
|
|
|
| |
This function can be used to create std::array without the need
to explicitly provide the size of array. It also has a specialization
that allow to generate sorted array at compile time. Sorted array can
be beneficial for example in binary search.
Change-Id: Ifc7e06e451812fce2ab94293959db5e9cc038793
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Apply a change from Qt Designer.
Silence clang-tidy warnings in the generated files
- use nullptr [modernize-use-nullptr]:
Use nullptr where applicable and simplify the switch()
statement in write() on this occasion to use member variables
instead of accessors.
- use '= default' to define a trivial destructor [modernize-use-equals-default]
Check if any delete statements are required. If not, use "= default".
- use auto when initializing with new to avoid duplicating the type name [modernize-use-auto]
Use auto.
Change-Id: I7b48872fd42a6118e7cca8510aacb9ffa7e01839
Reviewed-by: Jarek Kobus <jaroslaw.kobus@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add a framework for reading and decoding stored graphical texture file
formats. Includes decoders for the PKM and KTX formats.
This is basically the same texture file reading that was added to
qtdeclarative for 5.11, but has been refactored to be independent of
the scenegraph and opengl.
Task-number: QTBUG-67026
Change-Id: I87d8117550d8a2112f4f58c03e9ac6b3249cbc5a
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
|
|
|
|
|
|
|
|
| |
ObjectBoundingMode has inconsistent behavior and is deprecated in
favor of ObjectMode.
Change-Id: I748f6283f3db5869bb9a67c08bf5f16abc6f95b0
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
|
|
|
|
|
|
|
|
|
| |
The enum class doesn't get treated as an int automatically, the
way that a plain enum does; so there's a silly compile error which
this cast fixes.
Change-Id: I21b56337e4f724e4de1819e22bc93e9af23c51ea
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
| |
While it is still formally correct, it does not add much
value anymore.
Change-Id: I32431e3e73f3ca662dc4beb754f53449692c56a9
Reviewed-by: Albert Astals Cid <albert.astals.cid@kdab.com>
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
|
|
|
|
|
|
|
| |
msvc2013 is not supported any more.
Change-Id: Idca32f77e1b97a219d8bb4571cfd41e7334ac84e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
| |
Change-Id: I351635c7cc0bb8c852626dc3c0a3caf33b25a761
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@qt.io>
|
|
|
|
|
|
|
|
|
| |
Now all files that use QInternalMimeData include
qinternalmimedata_p.h, and we can remove the transitioning include
from qdnd_p.h
Change-Id: I6c86cb6a5bd73076e5dd3bafc01861b75ab83845
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
|
|
|
|
|
| |
Change-Id: Id59bdd8f1a804b809e22fffd154078f047078d70
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The pseudo classes left and right for QTabBar were mixed up. This
resulted in the fact that the east (=right) tab position was selected
with QTabBar::tab::right and the west with ::left
[ChangeLog][QTabBar][QTabBar] fixed mixed up css pseudo class for
left and right
Task-number: QTBUG-18146
Change-Id: I9f485f21c0a1c54bfac757a6f530aeeeef9b08bc
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A RAII style class which calls a function at end of scope.
Example usage:
auto cleanup = qScopeGuard([] { <my cleanup code> ; });
[ChangeLog][QtCore] Introduced QScopeGuard.
Task-number: QTBUG-62894
Change-Id: Ife67f5c76255a1fafbae03367263da0bac9a0070
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|
|
|
|
|
|
| |
The flag is deprecated.
Change-Id: Ice6ffafac6a5fa79675fc7cd84a9fb746f17dfb5
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
| |
The flag is deprecated.
Change-Id: Idfd7c31278032ee96b27f3f447a97ecfdd8051af
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
| |
Change-Id: I72d597fa21521a04b7f7c0e41bd45ee9dabb6222
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
To quote the documentation:
Queries and returns the state of the modifier keys on the keyboard.
Unlike keyboardModifiers, this method returns the actual keys held
on the input device at the time of calling the method.
So GetCurrentKeyModifiers seems to be the correct Carbon
function to use.
[ChangeLog][QtGui][QGuiApplication] Fix queryKeyboardModifiers() on
macOS to actually return the current modifier key state
Change-Id: I11f2ef1897a39aea13df4afbfebb8172ca803a30
Task-number: QTBUG-26413
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
|
|
|
|
|
|
|
|
|
|
| |
As remote based printers can be slow to query, meaning that it will take
time to check all the capabilities, especially if the QPrinter is
recreated on need. Then the information should be cached and reused if
the printer is deemed to be the same.
Change-Id: If8f9626c0d44113c698b4f61bbd197157932ccbe
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
|
|
|
|
|
|
| |
Change-Id: Ie8c3975643b9d7e60caf9de4dd3eaa13066a1028
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When flushing parts of a QBackingStore onto child QWindows, and we're
in layer-backed mode, we need to set the contentsRect of the layer
so that the layer will only show the part of the top-level-owned
backingstore image that's relevant for the child window.
Since the contentsRect is in unit coordinate system, we need to apply
a transform, but this must be based on the top level view's size, not
the image size, as the latter will be twice as big on a retina screen,
giving the wrong unit rect.
Change-Id: I7d6f378ed46a98272efb13406a2878ec1efe734e
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The underlying QOpenGLContext may fail to be made current, e.g. if the
surface is not exposed, or the graphics hardware is not available.
Instead of trying to initialize and resize the GL viewport with a
non-current context, we return early and defer the init and resize
until later.
Change-Id: I278ca8f1ad4d3da2d5be18b44d775f8d6c8af726
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
|
|
|
|
|
|
|
|
|
|
| |
The macOS bug preventing us from doing this has now been fixed:
http://openradar.appspot.com/radar?id=4976602949615616
Change-Id: I3bfa75d6bf982a051a9b274f530529f4ec4351a0
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Since MySQL 8.0.1 the my_bool type has been removed because all the C
compilers supported by the project supports C99 which has bool through
<stdbool.h>. This patch makes the driver both forward and backward
compatible.
[ChangeLog][QtSQL][MySQL] Now supports MySQL versions >= 8.0.1.
Change-Id: Ib373992763643bd470d33e85ce5cec7e4ae5e895
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
|