There should be no blank lines in \value content, otherwise qdoc
ends the list and starts a new one.
Change-Id: Idddc7992317894487445aea36397136df40b9691
Reviewed-by: Martin Smith <martin.smith@qt.io>
C++17 adopts P0021R1[1], which makes noexcept be part of the function
pointer's type and thus be overloadable. It contains some provisions for
allowing a noexcept function pointer to cast implicitly to a non-
noexcept function pointer, but that fails in the presence of templates
and additional overloads that could match the type in question.
Fortunately, the paper proposed a test macro, so we can change our
sources now and be compatible with both C++14 and C++17 rules.
This first failed with Clang 4.0 trunk. This source incompatibility is
not our fault, it's the language's doing.
[1] http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2015/p0012r1.html
Task-number: QTBUG-58054
Change-Id: I2bc52f3c7a574209b213fffd14988cf0b875be63
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
It's supposed to be a full alignment (like Top | Right), not
just one flag.
Change-Id: I656adda83742d7e4f31955322e937e979b32747c
Reviewed-by: David Faure <david.faure@kdab.com>
For language "Traditional Chinese" on iOS with region "US",
the logic was formerly to attempt a match on country/language/script (fail),
followed by country/language (which would result in script defaulting to
"Simplified"). Now, the logic is to try language/script first if script is
specified. Failing that, language/country will be attempted.
Task-number: QTBUG-39639
Change-Id: I75a774b1e66686e95167ff221458a97a7ea2660d
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Jason Erb <jason.erb@sparist.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
If we use QLatin1String we should include qstring.h.
Change-Id: Iebd761b98e515e9cd9cd34b96a0f8a602d00f086
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Don't use Dinkum choices when the C++ library is libC++ (QNX 7.0).
Change-Id: I18c3f716ccfb0c02dbfdc01eac4b707d3ae9aab6
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
StripTrailingSlash removes trailing slashes from the path, but not the
entire URL.
Task-number: QTBUG-47607
Change-Id: Id62b971e563e290b7ca000576bcc328616a3f1a2
Reviewed-by: David Faure <david.faure@kdab.com>
turns out that just appending builtin-qt.conf isn't a good idea:
executable-editing tools (objcopy, prelink, etc.) will happily drop the
"attachment".
a safe method would be adding a proper section to the executable, but
there doesn't appear to be an objcopy equivalent in msvc, and using
entirely different methods of embedding the file with different
toolchains seems like a rather bad idea.
so instead go back to the old method of building qmake with a generated
qconfig.cpp. of course, as said file is now created by qmake itself, we
have to compile qlibraryinfo.cpp a second time, and link a second qmake
executable.
Task-number: QTBUG-57803
Change-Id: I9e232693550aa870cec154e49cc06add13017cc2
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
We have to enable qt_safe_ftok with either sharedmemory or
systemsemaphore. In order to make the resulting QT_CONFIG work with the
bootstrap library we switch the features off for bootstrapping. Some
tests and examples have to be excluded when sharedmemory is not
available.
Change-Id: I3fc3926d160202b378be2293fba40201a4bf50c5
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
In this case, the issue was ICC, when compiling QtQml:
qv4sequenceobject.obj : error LNK2019: unresolved external symbol "__declspec(dllimport) public: void __cdecl QList<class QItemSelectionRange>::replace(int,class QItemSelectionRange const &)" (__imp_?replace@?$QList@VQItemSelectionRange@@@@QEAAXHAEBVQItemSelectionRange@@@Z)
referenced in function "public: static bool __cdecl QV4::QQmlSequence<class QItemSelection>::deleteIndexedProperty(struct QV4::Managed *,unsigned int)" (?deleteIndexedProperty@?$QQmlSequence@VQItemSelection@@@QV4@@SA_NPEAUManaged@2@I@Z)
This applies the same fix as qvector.h has had for ages due to QPolygon.
Change-Id: I15b62e0f9cec482fbb40fffd1490d791db5056bc
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
getMaximum() and getMinimum(), called during parsing, create new
QDateTime instances, which on Linux end up calling mktime().
Making these static (for the common case of LocalTime spec)
improves performance dramatically, when parsing several date/times.
tests/benchmarks/corelib/tools/qdatetime/ (after fixing it to
actually parse a valid date/time) says:
RESULT : tst_QDateTime::fromString():
- 36,742,060 instruction reads per iteration (total: 36,742,060, iterations: 1)
+ 24,230,060 instruction reads per iteration (total: 24,230,060, iterations: 1)
Change-Id: I0c3931285475bf19a5be8cba1486ed07cbf5e134
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The article does answer the questions asked in the bug.
Task-number: QTBUG-56629
Change-Id: Ib8bac0acf45bc10598fc47feb6dd73005b5ad040
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Two sentences that read like author notes likely added by accident.
Task-number: QTBUG-56630
Change-Id: I7a0b114e128f95e54e9e8f26b43493f67747d650
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
This code always intended to follow the recommended checking order from
the spec, which says "if multiple globs match, use file contents".
But if the globs had different weights, it would discard globs with
lower weights, and then wrongly conclude, if there is only one glob
left, that there was no ambiguity.
The correct way is rather: remember that multiple globs matched,
do determination from contents, and if that didn't work, *then* use
(one of) the highest-weight glob(s).
This fixes PGP-encrypted *.asc files being detected as text/plain rather
than application/pgp-encrypted.
(https://bugs.kde.org/show_bug.cgi?id=346754)
Change-Id: I734459daf9f502baa95ebb89432819964e0ce304
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
* apply endianness to the mask as well
* do not apply endianness to "host" entries, they were wrongly behaving
exactly like "big endian" entries.
The issue with the mask was detected by the audio/aac magic
<match type="big16" value="0xFFF0" mask="0xFFF6" offset="0"/>
which failed to identify the test file ct_faac-adts.aac since it was
applying the mask 0xFFF6 instead of 0xF6FF (on a little-endian machine).
Not yet detected by tst_qmimedatabase which is based on shared-mime-info 1.0,
will be covered by the upgrade to 1.8 in dev.
Change-Id: I4fb7af2d367099817e712b14f2a031066d0ac432
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Since the loop says p <= e, no +1 should be added to e.
Testcase:
The magic for application/x-gameboy-rom says
<match type="byte" value="0x0" mask="0x80" offset="323"/>
and this code was checking both byte 323 and byte 324, finding a match
at pos 324, returning application/x-gameboy-rom erroneously.
Given the magic for application/x-gameboy-color-rom:
<match type="byte" value="0x80" mask="0x80" offset="323"/>
the expected result for game-boy-color-test.gbc is application/x-gameboy-color-rom
Not yet detected by tst_qmimedatabase which is based on shared-mime-info 1.0,
will be covered by the upgrade to 1.8.
Change-Id: I2396cb1ccfb26db5a24d5551fef493cc0b98a247
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
initially, the idea was that QLibraryInfo would receive a
pre-sysrootified ExtPrefix from the builtin qt.conf. matching this
against the sysroot would then tell us whether to sysrootify the Prefix
from a "regular" qt.conf as well.
however, this would have lead to some major ugliness and inconsistency
between the code paths, so i changed my mind.
unfortunately, i failed to adjust the remaining code, leading to
169a40d51 entirely breaking sysrootification ...
the proper (and nicely consistent) solution is to introduce a
SysrootifyPrefix key to qt.conf. this is user-accessible as well, so as
a bonus it is now possible to adjust the setting at qmake installation
time. incidentally, this omission was the last thing that prevented
using the same qmake host build for any imaginable configuration of the
same qt version ... i think.
Change-Id: Ic0eebf21f93651f6374628c0ad8b206d696a4a7e
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
We chose to use UTF-8 as it allows us to ensure there's no mistaking the
hostname in case the locale is changed, if the host name contains
characters outside of US-ASCII.
But this didn't work because the code that wrote the hostname always
used the local 8-bit codec instead of UTF-8. On Unix, we used the
result of gethostname(3) directly, which is supposedly on the locale codec.
This commit doesn't fix Windows, which requires _wgetenv, the plan being
to encapsulate that with a qEnvironmentVariable() method.
[ChangeLog][QtCore][QLockFile] Fixed a bug that caused QLockFile not to
recognize a stale lock file if the machine's hostname contained non-US-
ASCII characters, on Unix. A Windows fix is still pending.
Task-number: QTBUG-49640
Change-Id: Ib9d045544ff370ec901626658a84ec4e6575fe21
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Callers can persist a QModelIndex which was not persisted before in a
slot connected to the signal, and such a persisted index must be updated
in the course of the layoutChange.
Store the indexes to persist after emitting the signal.
Task-number: QTBUG-32981
Change-Id: Ibee4c0d84817d72603a03fe5b22fdeefeac0695e
Reviewed-by: David Faure <david.faure@kdab.com>
Similar to the fix in the parent commit, incorrect updating of the
internal data structures during layout changes can lead to dangling
pointers being dereferenced later. Moves are treated as layoutChanges
by this proxy by forwarding to the appropriate method. However, data is
incorrectly cleared prior to that forwarding. Remove that, and let the
layoutChange handling take appropriate action.
Change-Id: Iee951e37152328a4e6a5fb8e5385c32a2fe4c0bd
Reviewed-by: David Faure <david.faure@kdab.com>
In sourceLayoutAboutToBeChanged the source model update is ignored if
the affected parents are filtered out anyway. The same logic is
attempted in the sourceLayoutChanged slot, but there the early-return
logic is applied too late - the mapping is cleared before performing the
early-return. Because pointers into the mapping are used in the
internalPointer of QModelIndexes in this class, persistent indexes used
later will segfault when attempting to dereference it.
Additionally, if a parent becomes invalid as a result of the
layoutChange, it would be filtered out by the condition in the loop,
resulting in a different result in the comparison of emptiness of the
parents container.
Fix that by persisting the parent's container, and performing the test
for early-return before clearing the mapping.
Task-number: QTBUG-47711
Task-number: QTBUG-32981
Change-Id: If45e8a1c97d39454160f52041bc9ae7e337dce97
Reviewed-by: David Faure <david.faure@kdab.com>
The process of loading a plugin is examplified by the qLoadPlugin
function (though not all users of QFactoryLoader use this function, they
all do something similar):
const int index = loader->indexOf(key);
if (index != -1) {
QObject *factoryObject = loader->instance(index);
if (FactoryInterface *factory = qobject_cast<FactoryInterface *>(factoryObject))
if (PluginInterface *result = factory->create(key, std::forward<Args>(args)...))
return result;
}
QFactoryLoader::indexOf already locked the mutex, but not
QFactoryLoader::instance. This commit fixes that.
Note that calling the virtual create() in the plugin's factory is not
protected by the mutex. Each plugin's factory must be thread-safe and
also create an object that works on any thread too. It's also the
responsibility of the caller of qLoadPlugin to ensure that it's called
thread-safely.
Task-number: QTBUG-42855
Change-Id: I63e21df51c7448bc8b5ffffd148ebee33d4c47de
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Clang pre-3.4 didn't like this and it's used in Xcode 5.1 (which we need
to support for 5.8).
error: 'this' cannot be implicitly captured in this context
typename T::const_iterator i = c.begin(), e = c.end();
^
Task-number: QTBUG-57488
Change-Id: I63e21df51c7448bc8b5ffffd148e688d7c9b89d6
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
After all the check makes sense here. If a timer was removed as a result
of sendEvent and it was not at the end of the list the list is not
shrunk but the timer info's id is just set to INVALID_TIMER_ID.
Additionally the timer's object should be fetched before we unlock the
locker as timerIdToObject is changed in removeTimer and we might access
a nullptr if the timer has been removed.
Reverts c83ba01f7b
Task-number: QTBUG-56756
Change-Id: Ib1a04c02fbfcf4c939b4891d42f954dc9e87149e
(cherry picked from commit 8f2088db17)
Reviewed-by: Maurice Kalinowski <maurice.kalinowski@qt.io>
Some public functions in QString and QDebug are declared inside
Q_COMPILER_UNICODE_STRINGS. This commit defines it for QDoc, and
adds documentation for QDebug functions that are now visible to
QDoc.
Change-Id: Ia7f2501c1dc7b8244dcc3ce4adcd2019fdbffcb6
Reviewed-by: Martin Smith <martin.smith@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
this moves us another step towards the "outer" configure doing just
minimal bootstrapping of qmake.
a challenge here was that so far, qmake itself needed qconfig.cpp. this
was replaced by usage of a qt.conf file instead of compiled-in values.
however, to make the executable still self-contained, that qt.conf is
embedded into it (by simple appending of a fixed signature and the text
file).
the qmake with the embedded qt.conf is not used for the qt build itself,
which instead relies on the qt.conf in bin/ as before. however, due to
the missing built-in values, this file now needs to contain more
information than before. but except for a minimal version that is needed
to start up qmake/configure at all, that file is now also generated with
qmake. as some of the newly set up properties are subsequently used by
configure itself, qmake gains a (deliberately undocumented) function to
reload the qt.conf after it's fully populated.
unlike the old implementations, this one doesn't emit redundant qt.conf
entries which match the hard-coded fallbacks. omitting them leads to
leaner files which are more comprehensible.
Started-by: Paolo Angelelli <paolo.angelelli@qt.io>
Change-Id: I4526ef64b3c89d9851e10f83965fe479ed7f39f6
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
The builtins clzs and ctzs have been removed. Additionally they were
never proper internal GCC builtins and shouldn't have been used in a
constexpr function in the first place. This patch removes the assumption
that they exist when BMI is available, and let GCC fall back to using
__builtin_clz and __builtin_ctz.
Change-Id: I3e0b4e246098bb9ce6ede28b311948260ef881b9
Task-number: QTBUG-56813
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reinterpret_cast()ing a pointer to a suitably sized integer is not guaranteed
to always give the same result for the same pointer (!). Therefore the
resulting integers are not comparable in a meaningful way. std::less is
supposed to be used to compare arbitrary pointers, so use it.
(Hopefully and reasonably, under the hood std::less does exactly what we
were doing, so this isn't BiC.)
Change-Id: I9960b3d6e35657fe7a25b842054f5d338280e850
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Add a bool *ok out parameter to qt_normalizePathSegments() and return false
when ".." are left over for an absolute path, indicating an attempt to
change above root.
Factor out static helper qt_cleanPath() to be able to pass the return value
to QDir::cd() and return on failure from there.
Amends change 63f634322b, which did
not handle UNC paths.
Task-number: QTBUG-53712
Change-Id: I3e63a5dd0259306a0b99145348d815899582f78e
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Calling QDir::cleanPath() on "//server/path/.." resulted in "/".
Factor out a function to determine the root path part of an absolute
path for later use, and handle some special cases:
- Consider server name of "//server/path/.." as part of the prefix.
- Check on the root path for WinRT.
Task-number: QTBUG-53712
Change-Id: Ibddacf06212b6fc86fa74a5e4078df6cfd5b66f5
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Maurice Kalinowski <maurice.kalinowski@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Commit c35fef9d3b wasn't sufficient. The
problem is that there's a complex combination of libc headers (math.h),
C++ headers (cmath), which may be provided by three different sources on
Linux (glibc, gcc and ICC). On some combinations, the isnan macro leaks
from math.h or cmath and that's what the the commit above tried to fix.
On some other combinations, there's no macro but there's an ::isnan
function defined. When we do "using namespace std; return isnan(x);",
that causes a compilation error. This commit solves that by detecting
whether there is a macro defined.
error: more than one instance of overloaded function "isnan" matches the argument list
function "isnan(double)"
function "std::isnan(double)"
argument types are: (double)
Change-Id: Iaeecaffe26af4535b416fffd148bf71826541bdd
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The C++ standard says in [support.initlist.access]/1:
constexpr const E* begin() const noexcept;
Returns: A pointer to the beginning of the array. If size() == 0 the
values of begin() and end() are unspecified but they shall be
identical.
So we can't assume it's non-null. I didn't want to remove the Q_ASSERT,
so passing a non-null pointer to append() remains required. This patch
simply won't call append() if the initializer list is empty.
This was already tested, but the failure is with a compiler that is not
part of the Qt CI.
Task-number: QTBUG-57277
Change-Id: Iaeecaffe26af4535b416fffd1489806872b412ee
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Dmitry Shachnev
Tasuku Suzuki
Liang Qi
Kavindra Palaraja
Marc Mutz
Thiago Macieira
Paul Olav Tvete
Giuseppe D'Angelo
Jason Erb
Rafael Roquetto
Ulf Hermann
James McDonnell
Oswald Buddenhagen
Aleksey Lysenko
Frederik Schwarzer
David Faure
Stephen Kelly
Oliver Wolff
Topi Reinio
Alexander Volkov
Allan Sandfeld Jensen
Friedemann Kleint