summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
...
* QPainter::stroke() on raster engine would draw moveto's as linesGunnar Sletta2009-07-231-5/+17
| | | | | | | | | The reason being that there was an assumption that any non-curved path was a continous polyline. For paths with multiple subpaths in it we need to split this up into multiple strokePolygonCosmetic calls. Task-number: 257621 Reviewed-by: Kim Motoyoshi Kalland
* Diagonal dashes are moving when touching the clip boundary.Gunnar Sletta2009-07-231-1/+4
| | | | | | | | | We normally pad the clip rect with the size of the pen and miterlimit to avoid this, but this didn't handle the case where there was a long diagonal dash. We also need to multiply the padding with the longest dash. Reviewed-By: Tom Cooksey
* Implement clipping in the QPaintEngineEx::stroke() function.Gunnar Sletta2009-07-233-6/+11
| | | | | | | This is a huge impact on performance whenever this path is taken. Reviewed-By: Tom Cooksey
* Enabled setting of DESIGNABLE=false-properties using FormWindowCursor APIFriedemann Kleint2009-07-231-4/+24
| | | | | | | Regression breakage introduced by the PropertySheet::isEnabled handling in 4.5. Task-number: 253278 Reviewed-by: Jarek Kobus <jkobus@trolltech.com>
* Recognize .jui file format in lupdateEskil Abrahamsen Blomfeldt2009-07-231-0/+3
| | | | | | | | The translator had been installed for the jui format, but the extension wasn't recognized. Task-number: 258547 Reviewed-by: Gunnar
* Remove autotest for "Won't fix" bug 258462.Rohan McGovern2009-07-231-31/+0
|
* return when brush is NoBrush in DFBPaintEngineAnders Bakken2009-07-231-0/+4
| | | | Reviewed-by: Noam Rosenthal <noam.rosenthal@nokia.com>
* Fix a bug in directfb mouse handlingAnders Bakken2009-07-222-15/+1
| | | | | | | | | | | Always check the mouse data we read for coordinates rather than asking the layer. This approach works whether we have layers enabled/window management enabled or any combination thereof. Also, make sure we use a software cursor when either of NO_WM or NO_LAYER is defined. Reviewed: Donald <qt-info@nokia.com>
* Use BatchBlit in flush/exposeRegionAnders Bakken2009-07-222-35/+47
| | | | | | | | | Minor optimization. Also make sure cursor is drawn in flush even if we're not in Offscreen mode. Reviewed-by: TrustMe
* Use dfbsurface::FillRectangles in solidFillAnders Bakken2009-07-221-5/+15
| | | | | | Minor optimization Reviewed-by: Donald <qt-info@nokia.com>
* Fix IDirectFBSurface::ReleaseSource callsAnders Bakken2009-07-222-2/+3
| | | | | | Make sure that these calls are in the right order. Reviewed-by: Donald <qt-info@nokia.com>
* Fix dfbwindowsurface handling for offscreen modeAnders Bakken2009-07-222-127/+151
| | | | | | | | | | | | | | This patch vastly simplifies the geometry handling (setGeometry/move) It also implements a mode in which DirectFB implementations that do not support windows can use an offscreen buffer as its backing store. Previously the only way to do this was to paint directly on the primary surface. This didn't work when the dfb driver didn't support an accelerated mouse cursor. It also detects the situation when the cursor isn't accelerated and takes care of painting it manually when needed. Reviewed-by: Donald <qt-info@nokia.com>
* Rewrite QDFBScreen::exposeRegionAnders Bakken2009-07-222-176/+124
| | | | | | | | | | | | | | | This code should be a noop in the case where one has a proper dfb cursor and proper dfb window handling. This was the only case it actually worked. This patch makes it work for a screen cursor rendered by Qt and sets exposeRegion up to work for using an offscreen backing store. Since one can't query the background color set by directfb this patch also adds a connect option to set the background color. This is needed for erasing the background when the mouse cursor moves. Reviewed-by: Donald <qt-info@nokia.com>
* Fix handling of invalid object paths and signatures in release mode.Thiago Macieira2009-07-221-11/+2
| | | | | | | | | | | | | | | | I had this #ifdef __OPTIMIZE__ there so that the compiler would know not to generate unnecessary calls and a long jump table for the switch of the marshalling code. Turns out that in release mode, the checks I added to make sure we detect invalid object paths and signatures were never hit (we always treated them as pure strings). So use the signature- and object path-checking code in both release and debug mode. Task-number: reported via email (tst_qdbusmarshall failing) Reviewed-by: Peter Hartmann
* Revert "Added a check that X11 timestamp goes forward only."Denis Dzyubenko2009-07-221-20/+15
| | | | | | | | | | | | | | | In some cases we might get an invalid timestamp that is far away in the future, so remembering it will break all consequent X calls that require a timestamp because it just contains junk (for example clipboard will stop working). This happens with XIM+SCIM pair - whenever we start input method and type something to the widget, we get a XKeyPress event with a commited string, however the 'serial' and 'time' members of the XEvent structure are not initialized (according to valgrind) and contain junk. This reverts commit 2ed015b8a0ffad63f0f59b0e2255057f416895fb. Reviewed-By: Brad
* QSslSocket autotest: adapt to new certificate on test serverPeter Hartmann2009-07-222-23/+16
| | | | | | got a new certificate, which is self-signed now Reviewed-by: Thiago
* Fix table borders in multiline tables when printing to PostScript.Gunnar Sletta2009-07-221-2/+7
| | | | | | | | | | | | | I'm not all to happy with this fix, but its the best that one can acheive given the current design. The problem is that QPdfBaseEngine sets a number of states as part of updateState(), but only when we are playing back through the alpha engine. These states are used in some draw functions, also when we are recording in the alpha engine. This leads to the states and their checks being out of sync. So to follow the existing pattern in the code we need to not touch d-> vars prior to a check to usesAlphaEngine. Reviewed-By: Eskil
* Fix memory leak.Olivier Goffart2009-07-221-0/+2
| | | | | | | | The signal could be connected a huge number of times This is already fixed in master with Qt:UniqueConnection Task-number: 258381
* Update documentation for QMessageBox::open().Norwegian Rock Cat2009-07-221-2/+4
| | | | | | I had missed this one in my rounds of updates. Bad me. Reviewed-by: Thorbjorn
* Designer: Fixed bug in setting QUrl property values from resources.Friedemann Kleint2009-07-221-2/+5
| | | | | | | Setting a file from a resource would result in 'qrc::/file' as the resource browser returns ':/file'. Reviewed-by: Jarek Kobus <jkobus@trolltech.com> Initial-patch-by: andy <qt-info@nokia.com>
* Fix a potential crash due to the fact that _q_UpdateIndex() is reenteredAndy Shaw2009-07-221-1/+1
| | | | | | | | | This is confirmed to resolve a number of problems from the original reportee. It's already fixed in Qt 4.6 in a more wider fix, but this one liner is a good to have in Qt 4.5.x anyway. Task-number: 258194 Reviewed-by: alexis
* Add a (failing) test for QProcess bug 258462.Rohan McGovern2009-07-221-0/+31
|
* Make sure DFB version macros are definedAnders Bakken2009-07-211-0/+1
| | | | Reviewed-by: Donald <qt-info@nokia.com>
* Another fix needed to build in a namespace on Mac with -arch ppcAndy Shaw2009-07-211-1/+2
| | | | | | | | Don't know how this got lost in the original submit since I had added both. Task-number: 257080 Reviewed-by: nrc
* QHttpNetworkConnection: Clarifying code comment about compressionMarkus Goetz2009-07-211-1/+6
| | | | Reviewed-by: TrustMe
* Fix building in a namespace when building with -arch ppc on Mac OS XAndy2009-07-211-2/+0
| | | | | Task-number: 257080 Reviewed-by: nrc
* QNAM: Proper loading of meta data when having AlwaysCache modeMarkus Goetz2009-07-212-0/+24
| | | | | | | | | Properly load the raw headers and properly handle the redirection when having a network cache in AlwaysCache mode (equals the offline mode in web browser). Task-number: 256240 Reviewed-by: Thiago Macieira
* Fixed compile on certain Solaris versions.Rohan McGovern2009-07-201-1/+1
| | | | | Every source file must end with a newline, otherwise: "Error: There is extra text on this line."
* s/slots/Q_SLOTS/Anders Bakken2009-07-202-2/+2
| | | | | | Fix QDirectFBMousePrivate and QDirectFBKeyboardPrivate Reviewed-by: TrustMe
* fix qmake syntaxOswald Buddenhagen2009-07-201-2/+2
| | | | Reviewed-by: TrustMe
* Fix QTextCodec case-insensitive comparison while in a Turkish locale.Thiago Macieira2009-07-201-4/+8
| | | | | | | In Turkish, lowercase('I') is 'ı', which means comparing "iso-8859-1" to "ISO-8859-1" will fail. Reviewed-by: Denis Dzyubenko
* Fix assert in message handling.Frans Englich2009-07-203-4/+28
| | | | | | | Trivial fix. Reported by Michael Brasser. Task-number: 258337 Reviewed-By: Peter Hartmann
* Fix the hand scrolling in QGraphicsView that will stop unexpectedly.Alexis Menard2009-07-201-1/+1
| | | | | | | | | | If you start a hand scrolling and during moving, you press another button of the mouse than the left one, the scrolling suddently stop working. In mouseReleaseEvent we just stop the hand scrolling if the button is left. Task:258356 Reviewed-by:janarve
* Fix crash when native socket notifiers would send a notification after being ↵Norwegian Rock Cat2009-07-201-4/+8
| | | | | | | | | | | | | | | | disabled. Spend a lot of time looking at this and at the CoreFoundation source code and it seems that we really do get a notification even after the notifier is disabled. I suspect there's a race condition between when we disable the socket notifier, but the kernel flags it as needing a read, then CoreFoundation just sends the notification without checking if the CFSocket has been disabled. No further notifications come of course. Since this breaks the invariant that was set in the assert, I'm replacing it with an if check. Task-number: 258198 Reviewed-by: Bradley T. Hughes
* fix crash due to null pointer referencingAlex2009-07-201-1/+1
| | | | | | | | | | during application desctruction globalEngineCache is deleted as part of Q_GLOBAL_STATIC macro. Other instances of code that happen to use QRegex after the cache destruction will subsequently crash. Most common reason are other Q_GLOBAL_STATIC instances which happen to use QRegExp as part of their destructor. Reviewed-by: Rhys Weatherley
* don't rely on system codec when handling PO filesOswald Buddenhagen2009-07-175-8/+17
| | | | | - make -input-codec affect PO files, default to UTF-8 - add -output-codec for PO files, same default
* QDirIterator: Doc fixes and whitespace cleanupJoão Abecasis2009-07-171-19/+12
| | | | | | There is no QDirIterator::isValid() function. Reviewed-by: David Boddie
* Refactoring QDirIteratorPrivate::pushSubDirectoryJoão Abecasis2009-07-171-26/+15
| | | | | | | | | | | | | | pushSubDirectory was operating on nextFileInfo when it should really be using the path received as argument. This fixes an issue introduced when currentFilePath variable was removed, that was exposed in the auto-tests; fixes a regression introduced in 4.5.0 -- test case a couple of commits back. This also allows refactoring calling code and avoid repetition. Task-number: 258230 Reviewed-by: Olivier Goffart
* QDirIterator: reducing "randomness"João Abecasis2009-07-171-10/+3
| | | | | | | | | | | | The difference between a canonical and absolute paths is subtle, and not what QDirIterator is about. With this change, we still avoid loops generated by symbolic links but won't duplicate entries because of these differences. While at it, when avoiding loops with symbolic links, please don't mess with the next path! That only added inconsistency. Reviewed-by: Olivier Goffart
* QDirIterator was returning inconsistent dataJoão Abecasis2009-07-171-6/+1
| | | | | | | One less variable to maintain reduces the number of bugs and improves consistency. Reviewed-by: Olivier Goffart
* Test case for QDirIterator regression introduced in 4.5.0João Abecasis2009-07-171-10/+21
| | | | | Task-number: 258230 Reviewed-by: Olivier Goffart
* Fix deadlock in the QWS server when destroying lots of windowsTom Cooksey2009-07-172-3/+8
| | | | | | | | | | | | | | | | | | | First, don't call QWSWindowSurface::winId() in the destructor, as it will actually request a new id if there isn't already one around - which is a bit silly and highlighted the "real" bug. Second, make sure QWSDisplay::Data::takeId() asks for 1 new id before waiting for more ids to arrive. This is because waitForCreation() calls QWSServer::processEventQueue(). If the events in the queue cause takeId() to be called, QWSDisplay::Data::takeId() gets called recursively. Even though there will be a create 15 ids command in the queue, that will only allow 15 QWSDisplay::Data::takeId() calls to return. The 16th call to QWSDisplay::Data::takeId() on the stack will not be able to return because all the IDs have been taken and (because it has been called recursively) no new create id commands have been generated. So the 16th call to takeId() spins in waitForCreate(). Reviewed-by: Paul
* reorganize numerus form count normalizationOswald Buddenhagen2009-07-179-30/+51
| | | | | | replace implicit normalization of individual messages on file writeout with global normalization which is called by the command line tools. this should a) be faster and b) cover the most critical case: lrelease.
* QUdpSocket: Doc improvementMarkus Goetz2009-07-171-0/+3
| | | | | Task-number: 236891 Reviewed-By: David Boddie
* Prepare device in DFBPaintEngine::beginAnders Bakken2009-07-171-0/+2
| | | | | | | 86ea4dbb5a748491656d9621ecd58238bc3e3d82 accidentally took out this line. Reviewed-by: TrustMe
* Mark virtual functions as virtual in DFBPaintEngAnders Bakken2009-07-171-27/+26
| | | | | | Make the code easier to read. Reviewed-by: TrustMe
* Remove unused function in QDirectFBPaintEngineAnders Bakken2009-07-172-34/+0
| | | | | | | drawColorSpan is never called from anywhere so we might as well get rid of the code. Reviewed-by: TrustMe
* Don't create dfbsurface in video mem if systemonlyAnders Bakken2009-07-171-1/+3
| | | | | | | If DSCAPS_SYSTEMONLY is specified we shouldn't try to create the surface in video memory. Reviewed-by: TrustMe
* Clean up directfb bit flippingAnders Bakken2009-07-178-69/+66
| | | | | | | | | | | | | | DirectFB declares variables that are bit fields as enums. E.g. DFBSurfaceCapabilities caps; caps |= DSCAPS_LOCK; // doesn't compile in C++ Work around this problem by declaring operators for these operations. This greatly improves the readability of the code. Reviewed-by: TrustMe
* Fix accidental selection of popup items under the mouse in QComboBoxmae2009-07-171-0/+2
| | | | | | | | | | If the widget under mouse is hidden, Qt can generate a synthetic mouse move event which gets delivered to the already hidden widget. This can then result in the wrong item being selected. Workaround: in QListView, ignore mouse move events when the widget is hidden. Reviewed-by: Denis
generated by cgit v0.12 at 2025-03-26 06:57:31 (GMT)