summaryrefslogtreecommitdiffstats
path: root/src/corelib
Commit message (Collapse)AuthorAgeFilesLines
...
* Doc: Fix QUrl::isLocalFile documentationSergio Ahumada2012-07-211-1/+1
| | | | | | | | | | | This function was introduced in 4.7 by a2f797b52c4274a62a7cf1f0939aca1429afe211 but then reverted by 98e935eed5549e479f6666680aed1711dc42111c Task-number: QTBUG-21293 Change-Id: I6ec9c6696e5c85e44774184010eb9135ef7f3365 Reviewed-by: David Faure <faure@kde.org> Reviewed-by: Casper van Donderen <casper.vandonderen@nokia.com> (cherry picked from commit e2c27b3052fe49a19c4092759682bddf2c2a37f7)
* Disable QAtomic assembly code for incompatible compilers.Ian Dean2012-07-201-1/+1
| | | | | Change-Id: I1d0798f766ea82ff18bcec10af4d002dab9c18d6 Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
* Only forward-declare some std classes if QT_NO_STL is defined.Raphael Kubo da Costa2012-07-171-0/+2
| | | | | | | | | | | | | | | | | If that is not done, building with clang, libc++ and C++11 support does not work: libc++ defines the iterator tag classes in an inline namespace inside the std namespace, and forward-declaring those classes inside the std namespace itself in qiterator.h causes an ambiguity that clang solves by choosing the incomplete-declared class declared in the enclosing std namespace. This is probably the "clang build problem" reported in commit 19e2b4d6eb733d9fd4eb69c0622b699fd08a3203 to qtbase, which cannot be directly cherry-picked without breaking other classes due to QT_NO_STL still being present in Qt 4. Change-Id: I26e33915347d3f94ef2e10b68fca6b2099ec9491 Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
* doc: removed incorrect reference to Q_OS_X11Rohan McGovern2012-07-131-1/+1
| | | | | | | | It's Q_WS_X11, not Q_OS_X11. Change-Id: I1778f6d268ea0ce702e5eb451f7a9dc0e1b5e754 Reviewed-by: Laszlo Papp <lpapp@kde.org> Reviewed-by: Girish Ramakrishnan <girish.1.ramakrishnan@nokia.com>
* Add Mac OS 10.8 QSysInfo enumsMorten Sorvig2012-07-102-1/+5
| | | | | | | Change-Id: If52192c5dcd8ba55b7ae7d3e3c4aea75c32d0f99 Reviewed-by: James Turner <james.turner@kdab.com> Reviewed-by: Christoph Schleifenbaum <christoph.schleifenbaum@kdab.com> Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
* QNX: QtGlobal: compile as a C header againMarc Mutz2012-07-091-1/+3
| | | | | | | | | | | | <utility> is a C++ header, only execute the check for the Dinkumware standard library if we're compiling under a C++ compiler. This is a backport of Qt 5 change b65291bc42c7e071e38efdd9ee6c4683f29aca13. Change-Id: I12d5a2c5c003092127773d9c25623d012976c906 Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
* QNX: fix compilation with C++11 enabledMarc Mutz2012-07-081-0/+12
| | | | | | | | | | | | | | | | | | | | | | QNX uses GCC, but by default not libstdc++ as the standard library, but 'libcpp', a Dinkumware-derived implementation that doesn't sport many of the C++11 features, yet. Thus, the compiler detection sets Q_COMPILER_INITIALIZER_LIST, which is correct, in a way, but since we're lacking stdlib support the next #include <initializer_list> will fail. So, simply don't define Q_COMPILER_INITIALIZER_LIST if we're on QNX and detect a Dinkumware signature (taken from Boost.Config). This is a back-port of change Ieeb147251c2935517faba61f75d1580a9e1649c4 from Qt 5. Change-Id: Ic7cd11b19a01ae192f60aae33da5581ff8e4682c Reviewed-by: Sean Harmer <sean.harmer@kdab.com> Reviewed-by: Thomas McGuire <thomas.mcguire@kdab.com> Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
* [Blackberry] Fix event loop throughput issueRafael Roquetto2012-07-051-8/+25
| | | | | | | | | | | Currently, only one bps event can be handled per event loop iteration, bringing about a huge performance penalty. This patch fixes this issue by bulk processing events whenever they are available. cherry-picked from qt5/qtbase 40bfbf3a5b765d939a7eea999c49da5b16456ad1 Change-Id: I52cc82d93754c2f6acf66f5b0445e34a68d233bf Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
* Fix socket notifier registration on Blackberry.Rafael Roquetto2012-07-052-8/+52
| | | | | | | | | | | | This fixes a bug on QEventDispatcherBlackberry, where registerSocketNotifier() and unregisterSocketNotifier() wrongly assumed that a file descriptor was being watched exclusively by one QSocketNotifier, while in fact there can be more than one QSocketNotifier associated with a single file descriptor. cherry-picked from qt5/qtbase: c10bc4c37d07dd7a9180d1810a379e370608c574. Change-Id: Iad2de22f54038e101de744a91be56db00891aca1 Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
* Fix access to uninitialized pointerTobias Hunger2012-07-031-1/+1
| | | | | | | | | | | | The C++ standard says in 9.4..2 that the object expression is evaluated, so any compliant compiler may access d. So this syntax is a bug in this place. This is a backport of commit 88cf9402e336fddeb673c92f3c14da47a9f8450b from qtbase. Change-Id: I02a9a1b73f7c18e8342aafeae412759f78086fa6 Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
* Removing potential crashes from out-of-thread cleanup on SymbianMurray Read2012-06-212-4/+37
| | | | | | | | | | | | | | | | | | | | There have been some crashes seen on Symbian when its adoptedThreadMonitor attempts to clean up objects created in other, now dead, threads. Some of these objects simply can't be cleaned up properly outside of their original thread, so the thread has to be checked when they are cleaned up, and cleanup skipped in the wrong thread. For pthread created threads, we also have the ability to insert cleanup code during thread shutdown. This was used in the 4.7 implementation of QThread on Symbian, and is a better solution for pthread based adopted threads as it gives in-thread cleanup. So the appropriate pthread code is also used with changes to adoptedThreadMonitor so that it can run along side the pthread cleanup code. Change-Id: Iad8207879b1ece62e5cce85f26a616166aa22486 Reviewed-by: Juha Kukkonen <ext-juha.kukkonen@nokia.com> Reviewed-by: Pasi Pentikäinen <ext-pasi.a.pentikainen@nokia.com>
* Avoid crash cancelling timer in wrong thread on terminateMurray Read2012-06-191-1/+1
| | | | | | | | | | | | | | | | | | When the Symbian timer active object detects that it is cancelled in the wrong thread, it was issuing a qFatal. But that was an overreaction, as the situation is not necessarily fatal if the thread in which the timer lives is about to die. For example if QThread::terminate is used outside of the thread targeted. Instead a debug message should be issued so that if there is a crash, the programmer should get information as to why. The debug message, in this case, is issued with RDebug::Print to avoid the use of Qt's message output which itself was crashing in partially finished thread situations. Task-number: ou1cimx1#1006699 Change-Id: I7d192eb27ba85ba2ef4f80fec3d8479f0df8f90c Reviewed-by: Juha Kukkonen <ext-juha.kukkonen@nokia.com> Reviewed-by: Pasi Pentikäinen <ext-pasi.a.pentikainen@nokia.com>
* Fix qatomic inline asm for x32 ABI.Davide Pesavento2012-06-191-3/+3
| | | | | | | | | | Drop the 'q' suffix from x86_64 asm instructions. It's not needed, because the assembler can automatically determine the proper size based on the target, and it is in fact causing compilation failures on x32. Change-Id: Ie6ff3ddf352a63bc490acce97a3019ce2e48dc70 Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
* UCD 5.0: Apply Unicode Corrigendum #6Konstantin Ritt2012-06-131-788/+789
| | | | | | | | | | | | | | http://unicode.org/versions/corrigendum6.html: > in Unicode 5.0, the list of characters with the Bidi_Mirrored property > was made consistent for brackets and quotation marks, in preparation for > new constraints on bidi mirroring. However, after publication of > Unicode 5.0.0 it was discovered that this change adversely affected > several quotation mark characters in deployed data. Task-number: QTBUG-25169 Change-Id: I2cd6a70a69f4b2c37a3f191e05233854a16f10d1 Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
* Fix QFile append on windowsShane Kearns2012-06-092-2/+10
| | | | | | | | | | | | 1. make windows file engine seek to end on open to match unix/symbian 2. make qfile forcibly seek the file engine in case a custom file engine has the same bug. Task-number: QTBUG-25906 Change-Id: I62a26519a9401e547ff77bd8a71027cb361dc671 Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com> Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com> Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
* Mac: Remove "unsupported" warning for 10.8Morten Sorvig2012-06-081-1/+4
| | | | | | Change-Id: I4e4447af16ab1aaa6b7a6eecf2e55718e4fb69d4 Reviewed-by: James Turner <james.turner@kdab.com> Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
* Add OS version detection for windows 8Andreas Holzammer2012-05-242-0/+6
| | | | | | | | | | Based on consumer preview, the internal version is 6.2 backported-from: I9d6ff6c7614f46a20d489e8a8f4aefeb60c547f6 Change-Id: I543148c22cec622ad04e6bf436acd0e66cbf02d2 Reviewed-by: Shane Kearns <shane.kearns@accenture.com> Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
* Fix local variable is initialized but not referenced warningNicolas Arnaud-Cormos2012-05-231-6/+7
| | | | | | | | The pragma needs to enclose the whole function. This change is not necessary in Qt 5 since QMap was rewritten. Change-Id: I63bb94ec09b9cb3c499938cde9c7fb690039406e Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
* QRegExp: fix crashGiuseppe D'Angelo2012-05-231-0/+2
| | | | | | | | | | | | | | | Fixes a crash when invoking various QRegExp methods on an object *before* doing any match. For instance fixes: QRegExp re; re.matchedLength(); // crash Backport of commit 136c2bf18446f2bbe7052d638c29edbc0b8ef6bc from qtbase. Task-number: QTBUG-23352 Change-Id: I9c239ff790a139c7820ef1aeced89d31320ae6b0 Reviewed-by: Andy Shaw <andy.shaw@digia.com> Reviewed-by: Janne Anttila <janne.anttila@digia.com> Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
* QRegExp: fix autotest, fix usage of uninitialized valuesGiuseppe D'Angelo2012-05-211-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | A (probable) typo was causing the code dealing with anchors to use uninitialized values. This used to work by chance, but was indeed detected by Valgrind f.i. when running tst_qregexp -- the indexIn test on anc11 data reported: ==3015== Conditional jump or move depends on uninitialised value(s) ==3015== at 0x514B4EA: PeppeQt::QRegExpMatchState::testAnchor(int, int, int const*) (qregexp.cpp:1813) [...] ==3015== Uninitialised value was created by a stack allocation ==3015== at 0x514B3EB: PeppeQt::QRegExpMatchState::testAnchor(int, int, int const*) (qregexp.cpp:1803) Fixing the code also makes the aforementioned test to succeed, therefore the #if 0 sections can be droppped. Backport of commits 281771ee201e591d4f40a161b93c71914b1b38f2 and 1fe7e557d75962c9c79cc344037a02ed50369430 from qtbase. Change-Id: I4d6cffdae737def1a47e72b46e40979e0aee1719 Reviewed-by: Andy Shaw <andy.shaw@digia.com> Reviewed-by: hjk <qthjk@ovi.com> Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
* Fix cases where functions are called with a drive and no slashAndy Shaw2012-05-161-9/+9
| | | | | | | | | | | | | | | | When a file is specified on a path that includes a drive letter followed by a colon but no slash then it didn't always account for the fact that this refers to the current path on that drive. This fixes the problems in completeBaseName(), baseName() and path(). Tests are also added for these three cases and some others too. Task-number: QTBUG-25353 Change-Id: I47a197c6af066f532442ad269be57597ec61303a Reviewed-by: Irfan Omair <irfan.omair@gmail.com> Reviewed-by: Shane Kearns <shane.kearns@accenture.com> (cherry picked from commit cfb44c6528b2518274bf157388832d1d610ce0e4)
* QElfParser: fix type of sh_sizePino Toscano2012-05-162-3/+3
| | | | | | | | | | The type of the sh_size field of a section header is either Elf32_Word or Elf64_Xword, so the type used cannot be qelfword_t (always 32 bits) but qelfoff_t. Change-Id: Ia380b6823913fee7a96b39f742630ae3a9ca0cb8 Reviewed-by: Thiago Macieira <thiago.macieira@intel.com> Reviewed-by: Arvid Picciani <arvid.picciani@nokia.com> (backport of 77b179689ba37dc909778fdd00df2701f83a2868 from qtbase)
* fix composition (C/KC) of some sequencesKonstantin Ritt2012-05-111-2/+3
| | | | | | | | | | | | | an unassigned code point should be treated like a starter code point; thus, if we have met an unassigned code point in the currently processed sequence, we should start process the next one (but we know there are no composed character for an unassigned starter -> simply skip it) This patch doesn't apply to qtbase, commit 46b78113b22428e6f8540193fcf0e00591dbd724 solved the same issue in a different way. Change-Id: I781a26024878d8b83a182b0fd4e681be2a6d9a27 Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
* Fix atomics on SPARCThiago Macieira2012-05-031-8/+8
| | | | | | | | | | Inspired by https://bugreports.qt-project.org/secure/attachment/26020/qt_atomic_sparc64.patch Task-number: QTBUG-22479 Change-Id: Ie3275df96c639d6a75e05f70fe5745aeb34457f9 Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com> (cherry-picked from commit 01ee0985319a0f41be12c3c8ff3813a74b43835e)
* Fix crash with evaluation licenseSamuli Piippo2012-04-271-7/+2
| | | | | | | | | | With QWS, an infinite loop is formed at application startup when commercial evaluation license is used and QT_EVAL is defined. Change-Id: Ia8e9a02c95e7277f49cf4866bae2422e6cef7973 (cherry picked from commit bf3c3862b486bbc472117d8e11e03f621b11c064) Reviewed-by: Donald Carr <donald.carr@nokia.com> Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
* Fix QThread start failure due to bad thread name on SymbianJuha Kukkonen2012-04-251-1/+10
| | | | | | | | | | | | | | | | | | RThread::Create() deems a thread name to be invalid, if it contains any of the characters: "*", "?", ":" or character is outside 0x20 - 0x7e range. This matches to the logic in User::ValidateName() that is used by RThread::Create() to validate thread name. In addition, maximum thread name length is 80 character on Symbian. It was possible that thread name contained e.g. colon that caused RThread::Create() to fail with KErrBadName (-28). Fix ensures that thread name contains only allowed characters. Task-number: ou1cimx1#996187 Change-Id: Ie6dd8c60bfed4e2f6cc48607ff0ff940d9cdae8a Reviewed-by: Murray Read <ext-murray.2.read@nokia.com> Reviewed-by: Pasi Pentikäinen <ext-pasi.a.pentikainen@nokia.com>
* Integrate Blackberry Platform Services (BPS) with Qt event loop.Jeff Kehres2012-04-255-7/+388
| | | | | | | | | | | | | This ensures interoperability between the Blackbery C and C++ APIs and makes it easier to expose platform services in C++ that are exposed in BPS - since events from both APIs can be processed on the same thread. Backport of qtbase 0723e14699704c35d5d61fa7f5e9f3bdbb378afa Change-Id: Ie0d66e7f9b9984ad3acff61f01665a2b90ef40db Reviewed-by: Sean Harmer <sean.harmer@kdab.com> Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
* Fix IPv6 address returned from QUrl::hostShane Kearns2012-04-231-5/+4
| | | | | | | | | | | | When passing an IPv6 address through QNetworkProxyQuery, it is stored in a QUrl internally. There was a bug in QUrl where it strips the [] surrounding an IPv6 address only if they were present in the input, otherwise it added them. Now the behaviour is the same as Qt5 ([] are always stripped). Change-Id: I42e020ce30d18a4108f1dd4428809fed07991680 Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
* Update QSettings and Mac OS X documentation with App Store informationPasi Matilainen2012-04-231-0/+21
| | | | | | | | | | | Add information about changing the location of global Qt settings to QSettings documentation, and general information about submitting Qt apps to the Mac App Store to the "Qt for Mac OS X - Specific Issues" document. Task-number: QTBUG-16549 Change-Id: Ic0210e8f4d3c3f6369032abd5c1e214b4200b179 Reviewed-by: Casper van Donderen <casper.vandonderen@nokia.com>
* Moved qmljs_debug_arguments to QCoreApplicationRafael Roquetto2012-04-202-0/+34
| | | | | | | | | Enable the use of the QML debugger without QtGUI Based on Qt5's 9a096d9e. Change-Id: If97502bc5367e0faadfaf3bbe70b0c89ef2c75c2 Reviewed-by: Kai Koehne <kai.koehne@nokia.com> Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
* Fix build with QT_NO_WARNING_OUTPUTAndy Shaw2012-04-191-0/+4
| | | | | | | | | | | | | | When defining QT_NO_WARNING_OUTPUT then qWarning() is not available so it should be defined in the same manner as qDebug() in this context which is as QNoDebug. The cases in QtNetwork are changed to use qWarning("....") to avoid having to add in an extra include of QDebug just so that qWarning() is declared in the QT_NO_WARNING_OUTPUT case. Task-number: QTBUG-25308 Change-Id: I4960d8943e805697d4c05cc6988306e5c25fc2bb Reviewed-by: Girish Ramakrishnan <girish.1.ramakrishnan@nokia.com>
* Bump Qt version to 4.8.2aksalova2012-04-181-2/+2
| | | | | | | | Change-Id: Iecae324b1ae00c0d6fb6ed36a8f4a27f51a31d07 Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com> Reviewed-by: Simo Fält <simo.falt@nokia.com> Reviewed-by: Juha Kukkonen <ext-juha.kukkonen@nokia.com> Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
* Add new QSysInfo::symbianVersion() enumsPasi Pentikainen2012-04-132-3/+13
| | | | | | | | | | Add new enums for future Symbian platforms. Task-number: QT-4593 Change-Id: I23b616c6d62d9d56d2b8ec6ac7edb8f6a84211e8 Reviewed-by: Juha Kukkonen <ext-juha.kukkonen@nokia.com> Reviewed-by: Janne Anttila <janne.anttila@digia.com> Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
* Use value of LANG if LC_TIME is empty.Thomas Sondergaard2012-04-041-0/+2
| | | | | | | Task-number: QTBUG-24912 Change-Id: I68d91c02f5d82b7fc4cae0d2dedafb862a3b323b Reviewed-by: Denis Dzyubenko <denis.dzyubenko@nokia.com> Reviewed-by: John Layt <jlayt@kde.org>
* Fix corelib build for WEC7.Janne Anttila2012-04-041-1/+3
| | | | | | | | | | | | | | | | | Qt5 backported fix http://codereview.qt-project.org/17852 broke the Qt4.8 build for WEC7. The change works fine in Qt5 because Qt5 does not anymore have QT_MODULE macro in every header, including qfunctions_wince.h. See commit: 4ecf8279 Apparently implementations of QT_MODULE and c-language zlib are not compatible, because MSVC2008 gives a lot of errors like: error C2085: 'qt_wince_getenv_s' : not in formal parameter list Build break fixed by putting QT_MODULE macro inside #ifdef __cplusplus. Change-Id: Id71386867750fcc61ab2cb2ecc527edc37c943ba Reviewed-by: Andreas Holzammer <andreas.holzammer@kdab.com> Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
* Fix QFileSystemEngine::tempPath on Blackberry OSRafael Roquetto2012-04-011-0/+7
| | | | | | | | | | | Unlike Unix, Blackberry OS stores the location of the temporary directory into the TEMP environment variable. Change-Id: I31f50fc4b160e5594339c4fcd62613963cc1484f Reviewed-by: Shane Kearns <shane.kearns@accenture.com> Reviewed-by: João Abecasis <joao.abecasis@nokia.com> Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com> Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
* Fix regressions due to partial QSslSocket::peek fixShane Kearns2012-03-301-0/+5
| | | | | | | | | | The fix broke HTTPS transactions with chunked encoding. It also broke use of a QSslSocket in unencrypted mode where peek and read calls are mixed. See change 68b1d5c17aa38d5921bdade2b0e0cb67c6c90513. Change-Id: Ib115b3737b0e4217496f5def10aaaea3c6452ff8 Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
* Fix scopes on iconv related pro/pri filesRafael Roquetto2012-03-281-1/+1
| | | | | | | | | | | | Ensure corelib and the iconv configure test link against libiconv on QNX-based platforms. This now includes the Blackberry mkspecs which were previously excluded. cherry-picked from qt5/qtbase: b0b36cc53c6115e5ca30986ceaf01215d09fcb39 Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com> Change-Id: Iee5758d3b31e3827e69a1598a5794b3d5ba25b03 Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
* Give QThread threads better names on SymbianMurray Read2012-03-271-6/+17
| | | | | | | | | | | | | | | | | | | | | | | | | | | QThread sets the name of a Symbian thread that it creates. This name can appear in a number of debugging scenarios, and it can give useful information. However the existing name set by QThread has not been informative. This change improves the amount of information in the thread name. Threads are now named as: <object name>_<class name>_v=<class vptr>_<random> The class name and vptr can be used to identify the QThread subclass type. In the case of a QThread subclass that generates a .moc file, the class name will be the true class name. However for other cases, eg where QThread is used directly, or is subclassed without a QOBJECT declaration, the name may simply be "QThread", and the vptr field might then be more informative. This change also allows the use of RTTI to get the true class name, through a rebuild of Qt, defining the macro name QT_USE_RTTI_IN_THREAD_CLASSNAME. This is not enabled by default, as there may be Symbian cases where RTTI does not work. Task-number: QTBUG-24950 Change-Id: Ifae9c6cb64638e95a01d34e421aa36f6fd0e7444 Reviewed-by: Shane Kearns <shane.kearns@accenture.com> Reviewed-by: Gareth Stockwell <ext-gareth.stockwell@nokia.com>
* Removing Symbian fast allocator implementationMurray Read2012-03-1610-6407/+2
| | | | | | | | | | | | | | An implementation of the Symbian fast allocator was added to Qt before it was available in all Symbian devices, to allow Qt the benefit of it in all installations. Now it is available in all Symbian devices and there is no need to keep a copy in Qt. So it is being removed to simplify the code. Task-number: ou1cimx1#946699 Change-Id: I937faca61beb4ef66a858766ebfa3deff564c786 Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
* Crash in Symbian text codec creation without CleanupStackMurray Read2012-03-151-1/+5
| | | | | | | | | | | | | | The Symbian text codec is creating a CCnvCharacterSetConverter with its NewL() function. This requires the CleanupStack, otherwise you get a E32USER-CBase 69 panic. But sometimes apps have been using text streams and other objects in static data construction or destruction which use the text codec where there is no CleanupStack. This change detects is a CleanupStack is present and temporarily creates one if not. Change-Id: Ia68fa60b4f0e3d675ee302208a625c6e9493a6f1 Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
* Fixes potential memory leak in qtextcodec.cppNick Ratelle2012-03-141-17/+9
| | | | | | | | cherry-picked from qt5/qtbase 4560ea230f40 Reviewed-by: Olivier Goffart <ogoffart@woboq.com> Change-Id: I06fc265fcc34bcc671c44b7041c0c461e588d193 Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
* Fix q_atomic_decrement for 64-bit SolarisSamuli Piippo2012-03-141-0/+2
| | | | | | | | | | | | | | | | q_atomic_decrement and q_atomic_increment did not set return value correctly. They expected the %eax return value register to be empty, when it was not necessarily so. This could occur when when compiler inlined e.g., ~QByteArray -> QBasicAtomicInt::deref -> q_atomic_decrement chain to application code. Data would not to be freed and leak memory. Fixed so that %eax is now zero'd before setting return value. Task-number: QTBUG-24470 Change-Id: I83495e3ddb938713863a7b00714215c331b9562b Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
* Fix documentation errorShane Kearns2012-03-131-1/+0
| | | | | | | | The function wasn't introduced in 4.8, we just documented its behaviour properly in that release. Change-Id: I0560be978612eaed4e632ea191c4b2ba19b4e7a8 Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
* Optimise QFile::seek that doesn't move the file pointerShane Kearns2012-03-131-0/+3
| | | | | | | | | If the seek position matches the current file position, then avoid doing an implicit flush that could be expensive. Change-Id: I019858ffb36fe832b9aee8da0a7803cafc8f7f75 Reviewed-by: Jaakko Helanti <ext-jaakko.helanti@nokia.com> Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
* Support changing locales at runtime in QPA plugins.Nick Ratelle2012-03-121-2/+33
| | | | | Change-Id: Id65798b81db2fa9fb5b1d929e4a94103995c6707 Reviewed-by: Jørgen Lind <jorgen.lind@nokia.com>
* Fixes mismatching delete operatorRafael Roquetto2012-03-111-1/+1
| | | | | | | | cherry-picked from qt5/qtbase 657f634c0e8a1bf4f544b Change-Id: I35cbdd6fa947a98a6afa70bdc72631c4213c288a Reviewed-by: Kevin Krammer <kevin.krammer@kdab.com> Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
* Fixed wrong iconv_open call under QNXRafael Roquetto2012-03-111-1/+1
| | | | | | | | | | | The wrong macro logic was causing iconv_open to never be called under QNX. cherry-picked from qt5/qtbase 8a15c41d36b1905a2 Change-Id: I0a1e3fb01d726daca637f8f51abfc924cd493311 Reviewed-by: Andreas Holzammer <andreas.holzammer@kdab.com> Reviewed-by: Harald Fernengel <harald.fernengel@nokia.com>
* Fixes codecs.pri to reflect new QNX mkspec nameRafael Roquetto2012-03-091-1/+1
| | | | | | Change-Id: I7d22d48b189e8abfd6164a51b1e025360ef3da5a Reviewed-by: Andreas Holzammer <andreas.holzammer@kdab.com> Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
* Enable storage of global Qt settings in app-local settings file on MacPasi Matilainen2012-03-072-2/+84
| | | | | | | | | | | | | The Mac App Store has limitations on where applications can write their settings, and com.trolltech.plist is not allowed. Changed the settings code to store all settings in the app-local file when the application runs in sandbox, or when the application's Info.plist contains a key "ForAppStore" with value "yes". The application's bundle identifier is also used for naming the settings file in these cases. Task-number: QTBUG-16549 Change-Id: Idd2241fbd7eb346da987226f05460642b0d6e5a3 Reviewed-by: Morten Johan Sørvig <morten.sorvig@nokia.com>
generated by cgit v0.12 at 2025-04-23 20:44:24 (GMT)