summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Check for null no longer neededJoão Abecasis2011-04-041-2/+0
| | | | | | | With proper checks for boundary conditions, we no longer need to protect against reaching the terminating null character. Reviewed-by: Robin Burchell
* Don't overrun XXXXXX placeholder when generating next filenameJoão Abecasis2011-04-041-3/+1
| | | | | | | | | | Checking of boundary condition needs to happen whenever there is a change. Checking anywhere else is useless. This fixes an issue where we would overrun the placeholder if there was a 'Z' next to the placeholder string. Reviewed-by: Robin Burchell
* Split argument validation of _gettemp as separate functionJoão Abecasis2011-04-041-25/+51
| | | | | | | | | This removes knowledge about the structure of the filename template from the randomization process and is a pre-requisite to adding Unicode support to the code. Task-number: QTBUG-4796 Reviewed-by: Robin Burchell
* Remove duplication with QCoreApplication::applicationIdJoão Abecasis2011-04-041-19/+5
| | | | | | | | | The duplication was there to allow qmake and other tools to compile without a full build of QtCore. Cleaning up this code justifies losing the benefit of PID-tagged temporary files in those tools. Reviewed-by: Olivier Goffart Reviewed-by: Robin Burchell
* Adding comments; variable scopingJoão Abecasis2011-04-041-22/+30
| | | | | | | There should be no semantic differences, otherwise. Reviewed-by: Harald Fernengel Reviewed-by: Robin Burchell
* Don't pre-check if directory exists as it's unnecessaryJoão Abecasis2011-04-041-35/+0
| | | | | | | | | | | | | | | | | | | On Posix platforms, when attempting to atomically create file and obtain descriptor, we only continue if we get an EEXIST error. If the path points to a non-existing directory, we should get ENOENT and will fail as well. On Windows, the check was already ignored. We would detect the file does not exist, attempt to obtain a handle upon return to QTemporaryFileEngine::open and fail there without running "for a *very* long time". Checking if the directory exists beforehand is not only unnecessary it also constitutes a race condition. Trusting the result of this check is dangerous and doubly useless. Reviewed-by: Olivier Goffart Reviewed-by: Robin Burchell
* QTemporaryFile: More dead code eliminationJoão Abecasis2011-04-041-84/+29
| | | | | | | | | | | | | | | | | | | There are exactly 2 uses of the _gettemp function: one for windows platforms, one for all others. Windows requests a file name, others request a valid file descriptor for that file name. Despite the 2 distinct use cases, the code was full of #ifdefs to keep never used code paths compiling. With this change the qt_mkstemp wrapper is eliminated, together with dead code inside if (doopen) {} blocks on Windows platforms. The more proper Q_OS_WIN macro is preferred to Q_WS_WIN. _gettemp is also changed to return the valid file descriptor on non-Windows platforms. Reviewed-by: Olivier Goffart Reviewed-by: Robin Burchell
* Removing dead code in QTemporaryFileJoão Abecasis2011-04-041-18/+4
| | | | | | | | The option to create a directory is never used, so let's drop it. There are bigger issues with this code. Reviewed-by: Olivier Goffart Reviewed-by: Robin Burchell
* Remove duplicate #includeJoão Abecasis2011-04-041-1/+0
|
* Extending fileTemplate autotestJoão Abecasis2011-04-041-14/+26
| | | | | | | | ... to also verify template prefix and corner cases with runs of less than 6 Xs. Reviewed-by: Olivier Goffart Reviewed-by: Robin Burchell
* Adding unicode test case for QTemporaryFileJoão Abecasis2011-04-041-0/+122
| | | | | | Task-number: QTBUG-4796 Reviewed-by: Olivier Goffart Reviewed-by: Robin Burchell
* Test should fail, debug output here is uselessJoão Abecasis2011-04-041-4/+1
| | | | | | Reviewed-by: Olivier Goffart Reviewed-by: Harald Fernengel Reviewed-by: Robin Burchell
* moc: qobject_cast cannot be used if the class privately iherit from QObjectOlivier Goffart2011-04-041-1/+3
| | | | | because the qobject_cast function do not have the right to access the staticMetaObject.
* Added debugging output to elf2e32_qtwrapper.axis2011-04-041-0/+31
| | | | Reviewed-by: Trust me
* Fixed detection of internal errors in the elf2e32_qtwrapper script.axis2011-04-041-6/+4
| | | | Reviewed-by: Trust me
* Added some more error handling to the elf2e32_qtwrapper.axis2011-04-011-4/+4
| | | | RevBy: Liang Qi
* QNetworkDiskCache: clean up after merge requestPeter Hartmann2011-04-014-9/+12
| | | | | | | changes include whitespace, parenting, recovering erroneously deleted code etc. Reviewed-by: Markus Goetz
* Speed benchmark for QNetworkDiskCacheSiddharth Mathur2011-04-013-1/+452
| | | | | Merge-Request: 2505 Reviewed-by: Peter Hartmann
* QNetworkDiskCache: change file organizationSiddharth Mathur2011-04-014-44/+78
| | | | | | | | | | Faster disk cache for mobile devices. Reduce file-system touching operations and work around FAT performance issues. Features new on-disk layout. Cached objects are now saved in multiple subdirectories and filenames are shorter in length. Merge-Request: 2505 Reviewed-by: Peter Hartmann
* Make sure we don't use subpixel positions in full hinting modeJiang Jiang2011-04-013-7/+4
| | | | | | | | | Some of the text can start from subpixel positions (start from 100.5 for instance), we need to make sure they stick to pixel grid in this case. Reviewed-by: Eskil (cherry picked from commit bf03b38d2b32a0cfca181398597884e4252466f4)
* Fix broken test on windowsOlivier Goffart2011-04-011-4/+0
| | | | | | | | | | | | | | stringbuilder.cpp(160) : error C2666: 'operator +' : 2 overloads have similar conversions stringbuilder.h(416): could be 'QStringBuilder<A,B> operator +<QByteArray,char[5]>(const A &,const B (&))' with [ A=QByteArray, B=char [5] ] or 'built-in C++ operator+(__int64, char [5])' while trying to match the argument list '(QByteArray, char [5])' The reason is the cast to QNoImplicitBoolCast (which is int)
* Fix compilation warningsOlivier Goffart2011-04-011-5/+4
|
* linking doc of QLocale with QApplication::keyboardInputLocale()Zeno Albisser2011-04-011-1/+5
| | | | | Reviewed-by: Denis Dzyubenko Task-number: QTBUG-17099
* Fix autotest when using QStringBuilderOlivier Goffart2011-04-012-25/+25
|
* Compile with QStringBuilder and QByteArrayOlivier Goffart2011-04-018-9/+10
| | | | Reviewed-by: Denis
* QStringBuilder now support building QByteArrayOlivier Goffart2011-04-014-34/+205
| | | | | | | | | This breaks source compatibility if one made its own QConcatenable as it nows require a Prefered type. And also sometimes if QT_USE_FAST_OPERATOR_PLUS was used, and the result of an addition between two QByteArray is used directly Reviewed-by: Denis
* Chnaged Q_AUTOTEST_EXPORT marco into Q_GUI_EXPORT for QZipWriter classPavel Mogilevskiy2011-04-011-1/+1
| | | | | | | | | Merge-request: 1162 Reviewed-by: João Abecasis <joao.abecasis@nokia.com> [ Symbols for QZipReader were previously exported with 959ab5d84d67ec84c2342b083e362ca75613e8cb. There's no good reason not to do the same for QZipWriter. Classes remain private API. -- João ]
* Fix horizontal centered text drawing without word wrapJiang Jiang2011-04-014-23/+30
| | | | | | | | | In this case we can't use line width to align the text, we have to use the rectangle width given in QPainter::drawText for alignment instead. Reviewed-by: Eskil (cherry picked from commit b92a31da01ee50fed4b10282e79f812167faf659)
* Fix bug when catching exceptions in tests.Marco Alanen2011-04-011-0/+2
| | | | | | | | | | Resets currentTestObject if an exception is caught while executing a test suite. This is required or else subsequent test executions will trigger an assert. Task-number: QTBUG-3869 Merge-request: 1163 Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
* fix order of #include-s on SPARCRitt Konstantin2011-04-011-2/+1
| | | | | | | | on some systems (i.e. on SPARC), the order of IPC-related includes is meaningful; sys/types.h, then sys/ipc.h, then sys/sem.h, and so on Merge-request: 2586 Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
* move semun declaration to qcore_unix_p.hRitt Konstantin2011-04-015-50/+27
| | | | | | | | | | | according to X/OPEN we have to define semun ourselves but on some systems sem.h will have it, so let's use a prefixed one (qt_semun). this allows to avoid the code duplication and makes things a bit easier to undesrtand Merge-request: 2586 Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
* Optimize QMutexPoolOlivier Goffart2011-03-312-13/+19
| | | | | | | Make the get() function inline and simplify the hash computation Reviewed-by: brad
* Fix declarative after the last changes to the QObject internalsOlivier Goffart2011-03-313-6/+6
|
* moc: be able to compile if there are private classesOlivier Goffart2011-03-314-10/+41
| | | | | Reviewed-by: brad Reviewed-by: thiago
* Speedup activation of signalsOlivier Goffart2011-03-3110-135/+221
| | | | | | | | | | | | | | | | | | | | | | | | | The virtual QObject::qt_metacall will recurse to the whole object hierarchy to find from which class a function should be called. But it is possible to know, at connection time, from which exact QMetaObject a function belongs, and the relative offset into it. So we make the slot calls from the qt_static_metacall function. So activation of signals is faster. - We must not call a slot from a class that has been destroyed. To avoid this, there is a check on the methodOffset. If it is smaller, that means we might be called (indirectly) from the destructor. We fallback to the virtual call to qt_metacall that does the right thing. - The signature of the static method is void (*) (QObject*,MetaCall,int,void**) It returns void, so the compiler is allowed to do tail recusive optimization. Having the QObject* as first parameter make it ready on the stack for the call to the member function. - The new static method has to be a member function in order to be able to access the private slots. Reviewed-by: brad
* Improve slightly the performence of activateOlivier Goffart2011-03-311-12/+12
| | | | | | | Avoid the overhead of QObjectConnectionListVector::at(), when we know anyway which list to take. Reviewed-by: brad
* Add a test in QObject.Olivier Goffart2011-03-311-1/+30
| | | | Testing that slot from destroyed class are not called
* Optimize QMetaObject::activate (remove call to QThreadData::current)Olivier Goffart2011-03-315-4/+9
| | | | | | | QThreadData::current is too slow, it needs to access the TLS The currentThreadId is much faster. Reviewed-by: Brad
* Remove NOPODINITIALIZATIONminiak2011-03-311-11/+0
| | | | | Merge-request: 1011 Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
* Replace CL_MIN/CL_MAX by qMin/qMaxminiak2011-03-311-23/+20
| | | | | Merge-request: 1011 Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
* Remove ANSI function definitionsminiak2011-03-311-4/+0
| | | | | Merge-request: 1011 Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
* Fix window pixmap initialization strategyJacopo De Simoi2011-03-311-3/+9
| | | | | | | | | | If a compositing manager is active, setting the window pixmap background to None for ARGB windows makes little sense. In this case the background should be transparent Merge-request: 1135 Reviewed-by: Denis Dzyubenko <denis.dzyubenko@nokia.com>
* Track compositingManagerRunning changesJacopo De Simoi2011-03-311-0/+9
| | | | | | | | | | | | | | | | Use XFixes extension to track changes to the compositing manager in such a way that QX11Info::isCompositingManagerRunning() returns a valid result even if compositing has changed after the creation of the QApplication Many thanks to Fredrikh Hoglund for suggesting the correct approach Task-number: QTBUG-6044 Merge-request: 1133 Reviewed-by: Denis Dzyubenko <denis.dzyubenko@nokia.com>
* QSslSocket: fix setReadBufferSizeMartin Petersson2011-03-312-14/+2
| | | | | | | This fix the qnetworkreply::ioPostToHttpsUploadProgress() auto test. Before the readbuffer where always limited to 1k for ssl sockets. Reviewed-by: Markus Goetz
* Merge remote branch 'origin/master' into 4.8-earthDenis Dzyubenko2011-03-3132-301/+510
|\
| * Merge branch 'master' of scm.dev.nokia.troll.no:qt/oslo-staging-1 into ↵Qt Continuous Integration System2011-03-2913-59/+144
| |\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | master-integration * 'master' of scm.dev.nokia.troll.no:qt/oslo-staging-1: Unix configure: Add support for --sysroot Add a 'capabilities' flag to the marshaller and demarshaller Add a set of connection capabilities to QDBusConnection Move the main D-Bus session and system connections to the main thread Add a way to obtain the DBusConnection* pointer from a QDBusConnection Make the QIconvCodec on Unix not enforce the Latin1 codec.
| | * Unix configure: Add support for --sysrootThiago Macieira2011-03-292-2/+27
| | | | | | | | | | | | Reviewed-By: axis
| | * Add a 'capabilities' flag to the marshaller and demarshallerThiago Macieira2011-03-297-34/+40
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The marshaller and demarshaller need to know the features negotiated with the peer so as to determine what's permitted. This is especially important to the marshaller: the libdbus-1 API may "throw a fit" if we try to pass a type that isn't allowed or the server may disconnect us. The use of the capabilities in the demarshaller are for symmetry and for us to toggle the feature in unit tests. Task-number: QTBUG-17478
| | * Add a set of connection capabilities to QDBusConnectionThiago Macieira2011-03-294-2/+31
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The capabilities are negotiated with the D-Bus peer or bus at connection time and may include extra features not available when D-Bus 1.0 was released. Currently (as of D-Bus 1.4), the only additional feature is Unix file descriptor passing. Proposed features are "maybe types" and single-precision floating point. They may be added to D-Bus 1.6. Task-number: QTBUG-17478
| | * Move the main D-Bus session and system connections to the main threadThiago Macieira2011-03-291-1/+11
| | | | | | | | | | | | | | | | | | Some applications experience weird behaviour that sometimes it works and sometimes it doesn't, due to D-Bus being first used in a thread. So instead do everything in the main thread for the two main connections.
generated by cgit v0.12 at 2025-01-09 15:11:36 (GMT)