summaryrefslogtreecommitdiffstats
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
* Diacritic input broken, duplicate the base glypha (on Mac OS X carbon)Prasanth Ullattil2011-06-091-2/+11
| | | | | | | | | This was a regression introduced by 8da880e77db04c. The saved key event needs to be cleared if we process a valid raw key event. Task-number: QTBUG-7099 Reviewed-by: Richard Moe Gustavsen (cherry picked from commit e30e3bb524536ef0f6b55272901b24bbbd4dfbd6)
* Fix QTransform::map(const QPainterPath &) not working with paths thatSimon Hausmann2011-06-091-1/+1
| | | | | | | | | | have only one element Replaced the isEmpty() check for the shortcut in map() with elementCount() == 0 Task-number: QTBUG-11264 Reviewed-by: Samuel (cherry picked from commit 90e009e9e273a4fea5166007e4c2b0638c6588cd)
* Styled item view backgrounds in Gtk+Jens Bache-Wiig2011-06-091-15/+41
| | | | | | | | | | | | These were currently never themed but used a fake gradient to mimick ClearLooks. I had to add a workaround for the case where you were using custom QItemDelegates since this could result in styled background on tree branches but flat items. Task-number: QTBUG-11209 Reviewed-by: thorbjorn (cherry picked from commit 29e294aee281b7a4dcdd1277eef0fe1552015536)
* QString: Fix severals bugs when comparing with QStringRefOlivier Goffart2011-06-091-4/+7
| | | | | | | | | | | | | | the internal ucstricmp and ucstrcmp contains different bugs if the strings are not 0-terminated, as it is with QStringRef. - in ucstricmp, even if the pointer are the same, the lenght could be different - we used to deference the 'end' pointer, that would be 0 if the string ends with 0, but we cannot do that in the general case Task-number: QTBUG-10404 Reviewed-by: Denis (cherry picked from commit cd003bfcf9a05967893099e8948ba3d8f281aa7d)
* QSslCertificate: support large serial numbersPeter Hartmann2011-06-091-4/+19
| | | | | | | | | | | We were calling an OpenSSL function that returned a long for the serial number; sometimes serial numbers are too big to fit into a long (up to 20 octets). In that case, do not convert the serial number to decimal, but just output the hexadecimal value. Reviewed-by: Zeno Albisser Task-number: QTBUG-9973 (cherry picked from commit 0f16c7ce8dcd6f4905d14875088c55148e41366a)
* Fixed alpha check for mono images.Kim Motoyoshi Kalland2011-06-091-0/+2
| | | | | | Task-number: QTBUG-9072 Reviewed-by: Trond (cherry picked from commit bcafe4a3dde9986ecfdeea43dc592b6f54134893)
* Fix antialiasing with transformed text in OpenGL2 paint engineEskil Abrahamsen Blomfeldt2011-06-095-4/+13
| | | | | | | | | | | | | | | | | | | | | | | | Since the paint engine now transforms the prerendered glyphs instead of rendering transformed glyphs as paths, we need to turn on texture filtering to avoid antialiasing artifacts. In order to do this, we also need to pad the glyphs in the glyph cache, otherwise you will get artifacts when sampling the area around the glyph's bounding rect (where there might be other glyphs.) This done by adding a glyphPadding() function to the cache which returns the number of pixels to pad between each glyph. Updated: This also fixes a general issue where some combinations of font sizes and transformations will cause the engine to sample neighbouring glyphs, so this has been backported to Qt 4.6.x. Task-number: QTBUG-9706, QTBUG-11028 Reviewed-by: Tom Conflicts: src/gui/painting/qtextureglyphcache.cpp src/opengl/gl2paintengineex/qpaintengineex_opengl2.cpp (cherry picked from commit 21f5bf3a3030a393ba17ce7726b70fe853b5b608)
* Fixed a crash when creating QGtkStyle before QApplicationJens Bache-Wiig2011-06-092-2/+2
| | | | | | Task-number: QTBUG-10758 Reviewed-by: ogoffart (cherry picked from commit bc5e4da28c1ffbc0da1cbfcaa1e6236ff4c2d4e3)
* Dont crash when assigning the same input context twice.Denis Dzyubenko2011-06-092-4/+10
| | | | | | | | | | | | Added a guard check to return if the given input context is the same is we already have. Explicitly mention in the doc that we take ownership of the given input context object. Task-number: QTBUG-10780 Reviewed-by: Thomas Zander (cherry picked from commit a32c96e753c2f5a123e518a92762ec9c9ff3b0b7)
* Fix platform regression: Support QFont::setStretch() on CocoaEskil Abrahamsen Blomfeldt2011-06-092-6/+15
| | | | | | | | | QFont::setStretch() would previously have no effect on Mac OS X/Cocoa. Task-number: QTBUG-10585 Reviewed-by: Eskil Done-by: Andy (cherry picked from commit 8cbf269e14b372ee3174c2706dd0cc4df37eef1a)
* Fix a memory leak from QItemSelectionModel when changing the modelThierry Bastian2011-06-091-1/+5
| | | | | | Reviewed-By: ogoffart Task-Number: QTBUG-10806 (cherry picked from commit 28b894f65c5aba01d5ee80c20f617478faaee7e7)
* Optimize QPixmapIconEngine's pixmap() functionaavit2011-06-091-14/+21
| | | | | | | | Replace needlessly expensive cache key generation Task-number: QTBUG-9850 Reviewed-by: Trond (cherry picked from commit ad6dafee9be288bcef6b2c4b318b234d2995abff)
* Fix compiler warning in QT_REQUIRE_VERSIONOlivier Goffart2011-06-091-1/+1
| | | | | | | | warning: format not a string literal and no format arguments Task-number: QTBUG-8967 Reviewed-by: Gabriel (cherry picked from commit 814f7d3d607edacca091273302297b6b2674424f)
* Fixed the sizing of the dock area with fixed size dock widgetsThierry Bastian2011-06-092-31/+19
| | | | | | | | | | It could happen that a dock area could be too wide. The problem was that we didn't really care whether or not the tab bar was visible. We would always take its minimum sizehint into account. Task-Number: QTBUG-10391 Reviewed-By: gabi (cherry picked from commit 8f53bf1b99d32caaebe8cbd1d9bd3a7381821988)
* Don't leak QVistaHelper from QWizardBradley T. Hughes2011-06-091-1/+2
| | | | | | | | | | The default QObject constructor was not called, meaning the vista helper was never registered as a child of the QWizard (and therefore never deleted when the QWizard was destructed). Task-number: QTBUG-10203 Reviewed-by: olivier (cherry picked from commit 6bc32600d2367e78ddc39dd93694e01d4d75958d)
* Fix a bug in QGraphicsItem::scroll.Alexis Menard2011-06-091-3/+4
| | | | | | | | | | If the rect argument is null, well we should not do the scrolling with an null rectangle but rather with the boundingRect like the documentation says. Task-number:QTBUG-10400 Reviewed-by:janarve (cherry picked from commit 44144cf60e978f7d5d70aec49d114d57832a78c3)
* Fix rtl issues with sliders in GTK styleJens Bache-Wiig2011-06-091-7/+16
| | | | | | | | | | | Sliders that draw groove decorations did not invert appearance correctly in vertical or rtl cases. With this fix we will flip the rectangles of the respective decorations and tell gtk if we are using an rtl widget or not. Task-number: QTBUG-8986 Reviewed-yb: thorbjorn (cherry picked from commit d52cc1ded5f30a6da0d11b9c3f7de66b2d3e9ee9)
* Fix missing pressed state for scrollbars with QGtkStyleJens Bache-Wiig2011-06-091-0/+5
| | | | | | | | | Pushing this to 4.7 as it is an easy fix and highly noticable in the latest official release of Ubuntu. Task-number: QTBUG-10396 Reviewed-by: thorbjorn (cherry picked from commit 5c1fe0fc017e116b2643d2b8278f7fca6fec10a1)
* Secure Cookies should only be sent over secure connections.Robert Hogan2011-06-091-0/+3
| | | | | | | | | | | | | | | | | | | | | | | http://bugreports.qt.nokia.com/browse/QTBUG-9618 QtWebKit currently fails the following test: LayoutTests/http/tests/xmlhttprequest/cookies.html This is because QNetworkCookieJar::cookiesForUrl returns secure cookies even when the connection is not secure. A 'secure' cookie is set by response headers from a http server as follows: 'Set-Cookie: cookie-name=value; secure' Correct QNetworkCookieJar::cookiesForUrl to ignore secure cookies when the url in the request is not 'https:'. Task-number: QTBUG-9618 Merge-request: 2372 Reviewed-by: Peter Hartmann <peter.hartmann@nokia.com> (cherry picked from commit 483fdd017d9998c6d7f4a035ca615e15fbc97e6a)
* Fixed the sizehint for cols/rows in qtableviewThierry Bastian2011-06-091-4/+4
| | | | | | | | | It could return wrong values because we were not always checking the right cols/rows. Reviewed-By: gabi Task-Number: QTBUG-10169 (cherry picked from commit 8bde410a6d0b1510d182a365d769db182b1c07dd)
* Cocoa: Native QFileDialog has problems selecting name filterRichard Moe Gustavsen2011-06-091-1/+1
| | | | | | | | | | | | The problem occured if you told the dialog to not show name filter details, then tried to select a filter specifying those details. The patch makes sure that we accept a filter as found in the drop down list if it matches the specified filter with or without details Task-number: QTBUG-8439 Reviewed-by: msorvig (cherry picked from commit 09c6d182a70f25626e12a739a9b7c3db937f47ce)
* QTabBar: Widgets inside the tab bar where not properly laid out after moveTab()Gabriel de Dietrich2011-06-092-11/+6
| | | | | | | | | | | Only the leftmost tab was being correctly laid out after the move due to a mistake in QTabBarPrivate::layoutTab(). Auto-test included. Reviewed-by: Thierry Task-number: QTBUG-10052 (cherry picked from commit 4743831d128dfad4ac9fbafa6e7544dbe7fb7ed2)
* Fixes wrong QGraphicsView's viewport cursor reset.Yoann Lopes2011-06-091-5/+7
| | | | | | | | | The cursor was reset even when it had not been previously saved. Autotest included. Task-number: QTBUG-7438 Reviewed-by: bnilsen (cherry picked from commit ad0c6b17cde4a3ce0041d1a54b96ccb5a6792520)
* Fixed assert failure when trying to load invalid SVG file.Kim Motoyoshi Kalland2011-06-091-1/+8
| | | | | | Task-number: QTBUG-10022 Reviewed-by: Trond (cherry picked from commit 13732b74e07fe10e95f2601f6cf2d0be78467eaa)
* QScrollArea: Excessive scrolling in focusNextPrevChild()Gabriel de Dietrich2011-06-091-2/+2
| | | | | | | | | | | We were scrolling horizontally even though the widget already fit the visible area horizontally. And conversely for vertical scrolling. Improves commit f1e92b2fdf6b6fd8aca0d05176c647dd68e21baa. Reviewed-by: Thierry Task-number: QTBUG-9425 (cherry picked from commit fa175f7c09e6e664f347ac9804b1a7db1d997ef2)
* Safeguard ourselves against corrupt registry values for cleartype gammaGunnar Sletta2011-06-091-0/+5
| | | | | | Reviewed-by: Eskil Task: http://bugreports.qt.nokia.com/browse/QTBUG-7596 (cherry picked from commit aab30e9a3ec78c3a9b2998feeef275e973ef0ff2)
* Force the repaint during a window resize.Fabien Freling2011-06-091-2/+3
| | | | | | Reviewed-by: Morten Sorvig Task-number: QTBUG-6384 (cherry picked from commit 3aee7881eaa45d4cbc1d0a353772cbf44638dadc)
* Fixed the action geometry of menu bar in RTLThierry Bastian2011-06-091-6/+5
| | | | | | | | | It could happen (with MDI widgets) that the first actions would not be visible. Task-number: QTBUG-9560 Reviewed-by: gabi (cherry picked from commit 864ace14e12cedd8e8f78841249e5309eb0a0795)
* Moved item wouldn't follow mouse cursor when ItemIgnoresTransformations is setGabriel de Dietrich2011-06-091-1/+2
| | | | | | | | | | | | Since the introduction of transform properties in QGraphicsItem, QGraphicsItem::transform() is no longer enough to get the correct transform of an item. No auto-test as it involves mouse dragging. Reviewed-by: Yoann Task-number: QTBUG-9295 (cherry picked from commit 253b87180e0a6c5db0feaaed7e321139c4ff1643)
* Total makeover of SVG image readeraavit2011-06-092-67/+106
| | | | | | | | | | | Improved/fixed canRead(). Added svgz support. Implemented ClipRect, ScaledClipRect and BackgroundColor support. Avoid data copy when reading from memory. Improved support reading from sequential devices. Added svg and svgz files to the qimagereader autotests. Task-number: QTBUG-8227 and QTBUG-9053 Reviewed-by: Kim (cherry picked from commit 2fe059c863377befdcf65084a07a7f4841beef0d)
* Fix the bug where the ampersand symbol would be discarded.Fabien Freling2011-06-092-3/+7
| | | | | | | | | We are now checking that the QAction is not related to the About or Quit button on Mac OS X Task-number: QTBUG-7540 Reviewed-by: Morten Sorvig (cherry picked from commit c19cb39ddbfd93802bfe61ecaa4f737a1bd1c1bb)
* Fix wrong arguments order in a warning in QObject::moveToThreadOlivier Goffart2011-06-091-1/+1
| | | | | Task-number: QTBUG-8478 (cherry picked from commit b7f48eee301e973fcfae08dfd8997538b6dbe251)
* _close(fd) closes the associated handle and not the other way aroundJoão Abecasis2011-06-093-2/+22
| | | | | | | | | | | ... according to the online MSDN documentation. Hid the cachedFd member in private data under WinCE, since it's never used there. Task-number: QTBUG-9085 Reviewed-by: Zeno Albisser (cherry picked from commit 7986ab58b9a5d0828291c857d3ce86bfa1af4e6e)
* QFSFileEngine: don't look through NTFS junctionsRitt Konstantin2011-06-091-8/+2
| | | | | | | | | | | | | | | | Starting in 1216161584b730576c24fb128131838be1826b37, we started processing NTFS junctions as symbolic link. This reverts isSymlink() behavior back to former behavior and processes only symbolic junctions. The code path there worked just fine for directory symbolic links but lead to multiple issues for volume mount point junctions due to incompleteness. Task-number: QTBUG-7036, QTBUG-7384 Merge-request: 493 Reviewed-by: João Abecasis <joao@trolltech.com> (cherry picked from commit b4996ab8e3a26d45ee55593da281e4b798c04dd2)
* After showing modal windows, WM_LBUTTONUP for double click is ignored.Prasanth Ullattil2011-06-091-0/+4
| | | | | | | | | We should not ignore WM_xBUTTONUP messages after WM_xBUTTONDBLCLK. Treat these messages like WM_xBUTTONDOWN messages. Task-number: QTBUG-7172 Reviewed-by: Thierry (cherry picked from commit 3fbedcd085b3e89e3646bf49264001ee216c5852)
* build fix for mingwThierry Bastian2011-06-091-1/+0
| | | | | | | we need to define WIN32_LEAN_AND_MEAN to include both windows.h and winsock2.h (http://msdn.microsoft.com/en-us/library/ms737629(VS.85).aspx) (cherry picked from commit 5d65e3c264332262fb09b8fe6f08d284c088a61d)
* Fix QGraphicsEffect cache when the item is not visible by the view.Alexis Menard2011-06-092-2/+2
| | | | | | | | | We can't rely on the exposed rect since the item can be outside the view so we need to regenerate the complete pixmap. Task-number:QTBUG-8750 Reviewed-by:sroedal (cherry picked from commit fcb738161efc965e91c3d528c31ac77d3a05ac64)
* Do not create native window handle just because a parent has one.Denis Dzyubenko2011-06-091-1/+1
| | | | | | | | | | On X11 when creating a widget that has a parent, there is no reason to create a native window handle right away since we don't know yet if the window is going to be ever shown. Task-number: QTBUG-9215 Reviewed-by: Bradley T. Hughes (cherry picked from commit 1d506981bcc3d2b8aad67989fd7946a0ad826856)
* Removed an assert in the print dialog on windowsThierry Bastian2011-06-091-1/+1
| | | | | | | | | | This assert was checking that the parent widget for the print dialog was already created. That's useless since we immediately after call winId(). Task-number: QTBUG-8684 Reviewed-by: gabi (cherry picked from commit 6d6df0e2b63f0103ec1f19b4b856b106e295d618)
* Do not deliver the same key event multiple times in Cocoa.Denis Dzyubenko2011-06-093-28/+19
| | | | | | | | | | | | When QCocoaView receives a keyDown or keyUp event, we translate it to QKeyEvent and deliver it to qt widgets, propagating if necessary. However if none of Qt widgets accepted the event we should propagate it back to cocoa - to get a beep from it, or for example if Qt widget is embedded into a native widget. Task-number: related to QTBUG-6444 Task-number: QTCREATORBUG-698 Reviewed-by: Richard Moe Gustavsen (cherry picked from commit 90dfb5e5d8fc8cb841b0762cd88aa4b996c38312)
* Do not beep on Mac when pressing some keys.Denis Dzyubenko2011-06-094-4/+17
| | | | | | | | | | When a key event is delivered to widgets, we should accept the event if the widget reacted to it in any way - if either the current item has changed or if the widget scrolled. Task-number: QTBUG-6444 Reviewed-by: Prasanth (cherry picked from commit 77230f7ec9e78ec2d57629ea934ceb1d59eb391e)
* Fixed a potential crash in headerview when inserting a sectionThierry Bastian2011-06-091-6/+3
| | | | | | | | This could happen if a section was moved andthen hidden Task-number: QTBUG-8650 Reviewed-by: Andy Shaw (cherry picked from commit f9b0efc17962df74a67d81daca5814af93a5fb97)
* Fix a sizing issue of message box on windows Vista/7Thierry Bastian2011-06-091-11/+38
| | | | | | | | | The problem is that the font changes after the creation of the buttons and we were setting a fixed size to the datails button. Task-number: QTBUG-8381 Reviewed-by: gabi (cherry picked from commit ee379340fcc5ebf8625906f42540c0fb3577f9ac)
* Make sure the delete and backspace keys are ignored when keyb-searchingThierry Bastian2011-06-091-0/+2
| | | | | | Task-number: QTBUG-8270 Reviewed-by: gabi (cherry picked from commit b912cac358a0a84f72f727767772caad22062a45)
* QNetworkReply: Fix canReadLine()Markus Goetz2011-06-093-1/+16
| | | | | Reviewed-by: Peter Hartmann (cherry picked from commit dd6c5cad88a56fb3a342fe9d4fc3b113ffe3fd53)
* Fixed problem displaying half width character as full widthDaisuke Kameda2011-06-091-1/+1
| | | | | | | | | | | | | | | | | | | | | | There is the problem that Qt displays half width character (e.g. alphabet) as full width. This is reported in #QTBUG-1726. This occurs in the following fonts. - Monospace font - Selif font - IPA font (Ver.0301) - NSimSun In Cario and fontconfig, the FT_LOAD_IGNORE_GLOBAL_ADVANCE_WIDTH is specified as flag of FT_Load_Glyph function to solve this problem. According to the document of freetype, this flag exists to support buggy CJK fonts. So this flag is specified as default flag of QFontEngineFT. Merge-request: 2319 Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com> (cherry picked from commit f18e95849be93b86ff014147086fa538993abc21)
* Fixed rounded SVG rectangles when 'rx' or 'ry' is 0.Kim Motoyoshi Kalland2011-06-091-2/+2
| | | | | | Task-number: QTBUG-8264 Reviewed-by: Trond (cherry picked from commit fb27eba721b41cb1b8557ffff663f9747a3922db)
* Slow QTreeView::layout() with many (> 10000) childrenGabriel de Dietrich2011-06-092-15/+51
| | | | | | | | | | | | | | | When updating the QTreeViewItem::total field in layout(), we used to call QTreeViewPrivate::viewIndex() to get the parent item, which is O(n). We have now introduced 'parentItem' in QTreeViewItem wich makes this O(1), with a small penality when inserting and removing elements in QTreeViewPrivate::viewItems. The QTreeViewPrivate::checkViewItems() is left inside the code for further debugging. Reviewed-by: Olivier Task-number: QTBUG-8885 (cherry picked from commit 7baaea978200c82fdf65e3934cfa373edeca6524)
* Added clipping to the dashed stroke processor in the GL2 engine.Kim Motoyoshi Kalland2011-06-094-30/+105
| | | | | | | | | | Without clipping, the stroker could consume a huge amount of memory when scaling up cosmetic, dashed strokes. I also made QDashStroker clip more aggressively. Task-number: QTBUG-7832 Reviewed-by: Samuel (cherry picked from commit ee6c26115959f46621f5a45f3fbe4a49989fd095)
* Cocoa: Fix misaligned text between format changesEskil Abrahamsen Blomfeldt2011-06-091-1/+1
| | | | | | | | | | | | | | | | | | | | This reverts part of commit e008504b5ec34975e34adf3b1a2b7170d0e4dd38. The original commit aimed to round up all fractional numbers for the font to "improve appearance". Rounding up the last advance of a set of advances in a glyph layout, however, breaks text layout completely, since the layout of text will be different depending on how many text lines is used to represent it (since the last glyph in each line will have a higher advance than the others.) This was e.g. visible in Qt Creator when turning on whitespace visualization, as all spaces then created a format change, thus a new text line in the text layout, and hence an increase of X pixels in the positioning of the text after the space, where 0.0 < X < 1.0. I see no negative results of removing the rounding, but if there are any problems, it has to be fixed in the usage of the advances not in the advances themselves. Task-number: QTCREATORBUG-381, QTBUG-8791 Reviewed-by: mae (cherry picked from commit ef9e7168a8ed3e8149204cf6c604a38db35245c0)
generated by cgit v0.12 at 2024-11-18 11:26:55 (GMT)