While the user is entering the password, the string variable that
stores the value might have to reallocate its content from time
to time (when the string needs to grow beyond its current capacity).
When the reallocation happens, the old buffer is freed, but its
data is not zeroed-out. This means that a QLineEdit that serves as
a password input field might leak chunks of the password during
its lifetime, and the leaks will persist after its destruction.
Since the QLineEdit can not control the behavior of the QString
it uses to store the entered value, the only thing it can do is try
to make the reallocations rare.
This patch reserves the space for 30 characters for the string which
stores the QLineEdit value when said QLineEdit is used for password
input. This is enough to make sure no reallocation happens in
majority of cases as barely anyone uses passwords longer than 30
characters.
[ChangeLog][QtWidgets][QWidgetLineControl/security] Preallocate a buffer
for the string that contains the entered value when the QLineEdit serves
as a password input field to minimize reallocations.
Change-Id: I3e695db93e34c93335c3bf9dbcbac832fc18b62d
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
The more modern LC_BUILD_VERSION load command was introduced in the 10.13
SDK to unify the various versions of the LC_*_VERSION_MIN command. When
building with a deployment target of 10.14, the linker will use this
load command instead.
Change-Id: Ic3571fdbfdf4dfb9346128c6f6e75d1e06f86cd2
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
this considerably speeds up failures, as no doomed build is attempted,
and produces more reliable results, as no second lookup (which would be
subject to environment changes) is done any more during the build.
in principle, this also opens up possibilities like selecting specific
variants of dependencies, automatically extracting rpaths, etc.
qt_helper_lib.prf also needs to create fully resolved library names now.
Change-Id: I65f13564b635433030e40fa017427bbc72d1c130
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Make the decoder fail early to avoid spending time and memory on
attempting to decode a corrupt image file.
Change-Id: I598db817c387867a449040f5be5427c8b8746483
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
If the document is paged and contains an image spanning more than one
page, correctly set the y position for everything following that
image.
Change-Id: I1c584c7a907c1728c2965f1dc3fdc56069ab3172
Fixes: QTBUG-59886
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
The contents of a deleted QString can still remain in memory
and can be accessible by tools that read the raw process memory.
This means that a QLineEdit that serves as a password input field
can leak the password after it is destroyed.
With this patch, the contents of the m_text string member variable
will be zeroed-out before the m_text is destructed. This is done
only in the cases when the QLineEdit serves as a password field.
[ChangeLog][QtWidgets][QWidgetLineControl/security] Zero-out the string
that contains a password entered into the QLineEdit
Change-Id: I8f88f952244bf8a0399c14acf0869439ca0a60ca
Reviewed-by: Luca Beldi <v.ronin@yahoo.it>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Certain fonts with multiple styles have the same family name. When
loading these as application fonts we were not specific enough when
querying for the text metrics. This meant that e.g. the bold version in
a font family would get the metrics of the regular one.
Fixes: QTBUG-67273
Change-Id: Ic988d62cddde0a1f77ddcaf2891cadc21c9b31e6
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
The alignment of long long on 32-bit windows is only 32-bits, but we
should be safe to assume it is aligned at 64-bit.
Since QRgba64 is a public type, we can't fix the alignment there without
breaking ABI, so instead fix it where temporary buffers are allocated.
At the same time be consistent about using QRgba64 so we can switch to it
enforcing alignment in Qt6.
Change-Id: Ie15c305bc867c62a13df8eb2b1678e92174e1f97
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
no need in duplicating DTLS-specific cases.
Change-Id: I475c6fb53daa44d60a5054bf3acc8474355b2186
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
If the application gets suspended when a blocking queued connection is
made then it will cause the application to hang when being resumed.
Therefore a check is needed to still post the event to the other thread
but in a non blocking manner so that it does not cause a hang on return.
Fixes: QTBUG-72101
Change-Id: I6d53c97ed6d9d500559da2a9fd195226d1fc9905
Reviewed-by: BogDan Vatra <bogdan@kdab.com>
Commits that added this code:
7ce8491280 and
f005dee1b3
I guess this was a work-around, before we started bundling
XCB libs a year later in 21bd66e1ea
0.3.8 was released in 2011-04-26 and we are bundling 0.3.9, so
it is safe to remove the include work-around as well.
Change-Id: I5794f40d86e10ebdad984aa4b61311979aaadcf3
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
This was a regression from Qt4, due to not setting the window
group leader in WM_HINTS property. Qt4 did set this property.
It was not obvious that something was missing because of the
similarly named WM_CLIENT_LEADER property, which we do set.
Testing revealed that setting WM_CLIENT_LEADER is sufficient
on some DEs, e.g. KDE, Unity, but did not have the desired
effect on e.g Gnome and XFCE.
EWMH/ICCCM specs are known to be ambiguous, WM_CLIENT_LEADER
is only mentioned in the section on the session management, so
it is not surprising that there is this inconsistency between
various WMs.
This patch merely restores a lost WM hint. This solves the
reported issues on major desktops. If we support calling show()
on a dialog before its parent has been shown is still an open
question, which should be handled in QTBUG-72040.
Task-number: QTBUG-56829
Fixes: QTBUG-46626
Fixes: QTBUG-70756
Task-number: QTBUG-72040
Change-Id: Id2c575850e5f4f5af3e57963c577d33572e30b6e
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
... printed by QPainter when it gets a pixmap with one or both of
dimensions equal to 0 (and a nullptr as a paintEngine).
Task-number: QTBUG-71806
Change-Id: I978f56c843daab307042e34390fc33f338ce8cf5
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
It's a leftover from when the Cocoa plugin used QFontEngineMulti
exclusively. Nowadays QFontEngineMulti will strip out the highByte
before calling into the real engine. Left an assert just in case..
Change-Id: I6cb26d20a908d7c3aaf096297fca160805fdb85b
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Commit 17b73b0d2b regressed performance by grabbing
the entire screen and then copying the grab region.
Move back to grabbing a minimal rectangle instead.
Also handle multi-screen grabs at different device
pixel ratios.
Multi-screen grabbing is handled by grabbing each screen
individually, and then stitching together the result.
The returned pixmap is created using the highest DPR
available, upscaling lower resolution parts as needed.
Change-Id: I9d45c5f7ec7b342360b745f49ef2939dc588f40b
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The only reason why we bundled this library ~6 years ago was because
it was not available on distributions that we supported at the time,
but library was a hard dependency for XCB plugin. See:
2122e731ab
Later more and more projects started to depend on it (compose input
context plugin, libinput, mir, wayland). The configuration had become
too complex, because some projects used bundled and some used the
version from the system.
Having libxkbcommon in 3rdparty sources is not necessary anymore, after
RHEL 6.6 was removed from the list of supported platforms for Qt 5.12.
Ubuntu 16.04 - 0.5.0
Ubuntu 18.04 - 0.8.0
openSUSE 42.3 - 0.6.1
RHEL-7.4 - 0.7.1
This will also simplify further development, e.g. QTBUG-42181
Bumped the minimal required version 0.4.1 -> 0.5.0.
The patch also contains a code marked with "TRANSITION HACK", which
is temporary needed so we can update the dependent wayland module.
[ChangeLog][Third-Party Code] Removed xkbcommon from bundled sources.
This library is present on all supported platforms. The minimal required
version now is 0.5.0.
Task-number: QTBUG-65503
Change-Id: Iec50829bb6f8fbb19f3c4e4ad62e332beb837de5
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
The code was previously assuming font-smoothing was only used with
A32 font antialiasing, so the corresponding gamma-correction was not
performed.
Task-number: QTBUG-71075
Task-number: QTBUG-71946
Change-Id: I68d8304cf18638239d8bfac32c67333f16ccc7bd
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Also, add a note that explains that autoRepeat does not affect
QToolButton's autoRepeat property.
Change-Id: I9e95cef9e9d1b5ee6cb1114d0b9a9fad562db601
Fixes: QTBUG-48204
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
How to test:
* Have two printers
* Use lpoptions -d to set the default printer to be one and then the
other
* Use lpstat -d to check setting the default printer worked
* Use this simple test program and check the resulting values make
sense
qDebug() << "DefaultPrinter" << QPrinterInfo::defaultPrinter().printerName();
const QList<QPrinterInfo> list = QPrinterInfo::availablePrinters();
for(const QPrinterInfo &pi : list) {
qDebug() << pi.printerName() << pi.isDefault();
}
Fixes: QTBUG-70317
Change-Id: I535d11451c568630a374f5c37d8cac32cbb6d3ab
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
Instead of tying the linear-conversion to a specific function, we move it
to imageForGlyph and base it on the premise for needing it.
Change-Id: Ib8fc79ad419ef703abcb82785ac15d4c75fb98e6
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
The font smoothing helper in particular now takes into account whether
or not we're dealing with color glyphs, in which case we shouldn't (or
don't need to) smooth, and also makes sure the QFont::NoSubpixelAntialias
style strategy doesn't affect font smoothing when we're dealing with
non-subpixel-antialiased font smoothing, as on macOS 10.14.
Change-Id: Ibd477158629402c55cafec31576b6d9901d184cf
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
The setting is not relevant for modern macOS applications, and none of
the applications shipped with macOS today are affected by it.
The only code path in macOS that picks it up is +[NSFont initialize] in
the UIFoundation framework, storing it for later so that -[NSFont screenFont]
and -[NSFont screenFontWithRenderingMode:] can use it, but these APIs are
deprecated and we don't use them in Qt.
Other NSFont code paths will not hit these APIs unless screen font
substitution is enabled, something it hasn't been since OSX 10.7.
https://preview.tinyurl.com/yctpfnqp
Removing handling of this setting allows us to simplify the reasoning
for whether or not antialiasing and font smoothing is enabled for a given
engine.
Change-Id: Ie2809052a1a0815d9bddedd4a6236eb6c898f993
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
No need to return top level window if it was not created.
It means no platform resources have been allocated.
Events might not be delivered if the wrong window is returned.
Fixes: QTBUG-70544
Change-Id: I43462974f70871470f7b7490dc2b3c08846f77b1
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
It will avoid a creation of temporary object and avoid copying.
Change-Id: Ifae5f6f9e36bcb07f4bacc31f151f8adcfa621a3
Reviewed-by: Anton Kudryavtsev <antkudr@mail.ru>
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
The 'aa' argument doesn't unconditionally enabled antialiasing, it just overrides
the check that the pointSize is larger than the antialiasing threshold. If the
styleStrategy has QFont::NoAntialias we still end up without antialiasing.
Change-Id: I7130e7c68d883c2443756242e96790264f583b0f
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Makes for clearer code than looking at the glyph format.
Change-Id: Id6dd2a7851aac2a42cc27d9e2fb408ce9a5345d3
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
The way macOS does font smoothing has changed in Mojave, and we need to
take both this new algorithm into account, as well as support users who
set legacy preferences to revert back to subpixel font smoothing.
As a followup to this patch we will tweak some of the existing logic
to take the new font smoothing algorithm into account, so this is
just a first step.
Change-Id: If37014c18515f406b8bb8194c9df7a75c2eb10fc
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Removes assumptions about QImage format in a few places.
Change-Id: I515701be53190429a48956c31986fa0804806406
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
The operator T() function of QAppleRefCounted should be const so
that the underlying type can be accessed from const member functions
just like the naked underlying type could.
Change-Id: I0819c5795d28442a6ff4db2732e211b183574f9f
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
The function doesn't need the flags argument, nor does it need the
ctfont or fontdef arguments if it's a normal const member function.
It can also be used from QCoreTextFontEngine::stringToCMap(), instead
of duplicating the code. This was originally the case before b4aa5d97
which improved surrogate pair handling, but for some reason the change
introduced the duplicate code instead of just changing the arguments
in the function call slightly.
The use of 0xff000000 to skip certain glyphs looks dubious, and is
probably related to QFontEngineMulti's use of the high byte to
indicate which engine the glyph came from, but the multi engine
strips this away before calling out to the concrete engine so
it could potentially be removed in a later patch.
Change-Id: I6c693595616da1b69fdbe3d7a31e392a8443369d
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
It's only used in a single function (twice), so let's keep it
closer to the call site.
Change-Id: I7f8ceadc380171237eef3fa6b03ccd6bc89e99af
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
The leak has been there since the beginnings of Qt 5.0.
Change-Id: I238181dcc63cb4cf8a60b5c565b184d8278d0315
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
Keeping the fake parent window around in a created state means we're
wasting system resources such as native platform views/windows, and
makes it harder to debug the lifetime of these resources.
The fake window will be re-created if re-parented into at a later
point.
Change-Id: Ib82560e7e565af19d58afe121fd087669a6ffb95
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Instead of relying on the return value of malloc having the correct
alignment, use proper non-throwing new[] operators.
Change-Id: I06c6c619e21c848f3d184bdb7cef8c5589c1c7ab
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
First of all, we were using the wrong size variable: instead of the size
of the section, found in the section header, we were using the size of
each section entry in the section table. Since that's usually smaller,
we weren't hitting a problem.
Second, if the string table is the last thing in the file and there's
nothing else after it, not even padding, then offset + section_size can
be equal to the file size. In fact, the .shstrtab section is usually the
last one, as it contains the section names themselves, so it stands to
reason that it's the second to last thing written.
For generic linkers, the last data in the file is the section table
itself, so usually the file is larger by at least a kilobyte, which is
why we haven't hit this bug. It could only manifest as deciding that
certain specially-crafted but valid ELF files were invalid. I can't
think of a way to trick it into thinking an invalid ELF is valid.
That's another reason why this code needs to be rewritten with more
modern coding styles and actually using <elf.h>
Fixes: QTBUG-71443
Change-Id: I1bd327aeaf73421a8ec5fffd156162f2df5557b8
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Originally QAtomicPointer didn't have the three-operand version,
resulting in code like:
if (!atomic.testAndSetXxx(expected, newvalue))
expected = atomic.load();
The three-operand version gives us the current value of the atomic in
case the test failed and it's free in all architectures, unlike the
extra load.
I have to use testAndSetOrdered here because I need the failing load to
use the Acquire memory order, even though that has an extra Acquire for
the successful case we don't need. QAtomicPointer does not have
testAndSetReleaseAcquire.
Change-Id: I1bd327aeaf73421a8ec5fffd1560fe30d3bfd9b8
Reviewed-by: Romain Pokrzywka <romain.pokrzywka@gmail.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@qt.io>
Ivan Čukić
Oliver Wolff
Tor Arne Vestbø
Oswald Buddenhagen
Eirik Aavitsland
Lars Knoll
Samuel Gaist
Mårten Nordheim
Allan Sandfeld Jensen
Timur Pocheptsov
Andy Shaw
Gatis Paeglis
Morten Johan Sørvig
Mitch Curtis
Ville Voutilainen
Thomas Miller
Albert Astals Cid
Val Doroshchuk
Mikhail Svetkin
Kirill Burtsev
Friedemann Kleint
Kevin Funk
Thiago Macieira