summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* let generated flag control SQL generationMark Brand2011-03-298-42/+62
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously, if QSqlField's QVariant value had type "invalid", this caused the field to be omitted in data-changing SQL statements generated from QSqlRecord edit buffers. Complaints against this mechanism: - It precludes initializing value type to field type, which would otherwise seem sensible and useful, such as when using model.data() in contexts where the field type is not readily available. - QSqlField::clear() does initialize value type to match field type and so is incompatible with this mechanism. - Problems such as described in QTBUG-13211. - Unwanted distinction between "invalid" and "null" when mapping field values to SQL values. - QSqlField's generated flag already provides a mechanism for controlling SQL generation. These complaints are redressed here by replacing this mechanism with reliance on QSqlField's generated flag. The flag is initialized to false in new edit records and set to true when a value is set. Applications can still manipulate generated flags directly to control SQL generation. Generation of SELECT statements already used the generated flag and is unaffected by this change. Merge-request: 1114 Reviewed-by: Marius Storm-Olsen <marius.storm-olsen@nokia.com> Reviewed-by: Michael Goddard
* Add version attributes as per ODF specificationTomas Straupis2011-03-291-0/+2
| | | | | Merge-request: 1120 Reviewed-by: Marius Storm-Olsen <marius.storm-olsen@nokia.com>
* 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.
* Add a way to obtain the DBusConnection* pointer from a QDBusConnectionThiago Macieira2011-03-292-0/+17
| | | | | | | | | | This is internal API and the behaviour may change without notice. In the future, I could use a different implementation without libdbus-1, so this pointer may be something different. The use of DBusConnection is also entirely implementation-defined, so keeping this connection around is entirely unsupported. You have been warned.
* Make the QIconvCodec on Unix not enforce the Latin1 codec.Thiago Macieira2011-03-291-20/+18
| | | | | | | | | | | | | | | | | | | | | | Originally, the QIconvCodec created and disposed of the iconv_t structure at every turn. In Qt 4.5, we started saving it for later in a thread-local storage. We had to introduce a fix to make sure that we didn't create the iconv_t structure before the setlocale(LC_ALL, "") call was made, though: otherwise, we'd keep forever an iconv_t that pointed to the wrong encoding. So now simply restore the Qt 4.4 behaviour: create and dispose of the iconv_t structure if we're called before the QCoreApplication constructor is run or after the static destructors are run. Note: this means QIconvCodec will probably default to US-ASCII when run before QCoreApplication, not Latin 1. Non-ASCII characters (anything with the high bit set) will fail to convert. Task-number: QTBUG-15229? Reviewed-by: Denis Dzyubenko
* tst_qlocale: improve failure message in tst_QLocale::macDefaultLocaleRohan McGovern2011-03-281-1/+5
| | | | | | Use QVERIFY2 to output more details about this failure. Reviewed-by: Toby Tomkins
* Add a once-over unrolled fromLatin1 conversion (32 characters)Thiago Macieira2011-03-271-0/+46
| | | | | This was suggested by Matthias Kretz on my blog. It improves on GCC (one case up to 5%), but worsens on ICC (up to 5%).
* Add 16-byte loads of the Neon fromLatin1 functionsThiago Macieira2011-03-271-0/+68
|
* Fix warning about unused parameterThiago Macieira2011-03-271-1/+1
|
* Merge branch 'master' of scm.dev.nokia.troll.no:qt/qt-earth-staging into ↵Qt Continuous Integration System2011-03-254-4/+57
|\ | | | | | | | | | | | | | | master-integration * 'master' of scm.dev.nokia.troll.no:qt/qt-earth-staging: Skip failing tests QDir: Fix absolute paths when using file engines
| * Skip failing testsJoão Abecasis2011-03-252-0/+26
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | c7748b7838522ec38ec01423f1267acf1f163606 fixed a bug in QDir::absoluteFilePath that was masking failures in QtScript's v8 test suite. Those will have to be looked into separately. For now, skipping the tests. These are the uncovered failures, broken out by platform and test, and formatted as "test-name:line-number: actual-value / expected-value": * linux-icc_ubuntu1004 - math-min-max:50: Inf / -Inf - negate-zero:50: false / true - str-to-num:50: Inf / -Inf * linux-icc_ubuntu1004 & macx-g++_cocoa_osx106 - smi-negative-zero:50: -Inf / Inf * linux-icc_ubuntu1004 & macx-g++_cocoa & macx-g++_cocoa_osx106 & win32-msvc2005_windows_xp & win32-msvc2008_windows_xp - to-precision:50: 1.235e+27 / 1.234e+27 Reviewed-by: Kent Hansen
| * QDir: Fix absolute paths when using file enginesJoão Abecasis2011-03-182-4/+31
| | | | | | | | | | | | | | | | | | | | | | | | QFileSystemEntry doesn't know about paths as implemented by custom file engines, such as is the case with resource files. In such cases, we need to allow the engine to make the relative to absolute conversion. Expanded QDir's test case to ensure resource paths are considered in more places. Task-number: QTBUG-17921 Reviewed-by: Prasanth Ullattil
* | Merge branch 'master' of scm.dev.nokia.troll.no:qt/qt-releng-staging into ↵Qt Continuous Integration System2011-03-252-20/+0
|\ \ | | | | | | | | | | | | | | | | | | master-integration * 'master' of scm.dev.nokia.troll.no:qt/qt-releng-staging: Removed old autotest references from Symbian build files
| * | Removed old autotest references from Symbian build filesEckhart Koppen2011-03-242-20/+0
| | | | | | | | | | | | | | | | | | | | | Tests not in use anymore, removed from the layers and package definition files. Reviewed-by: TrustMe
* | | Merge branch 'qt-master-from-4.7' of ↵Qt Continuous Integration System2011-03-2417-85/+283
|\ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | scm.dev.nokia.troll.no:qt/qt-integration into master-integration * 'qt-master-from-4.7' of scm.dev.nokia.troll.no:qt/qt-integration: Icon in QMessageBox::question is deformed on Symbian Prevent crash when connecting signal in splitview GraphicsViews: Focused widget is not visible after orientation change GraphicsViews: Focused widget is not visible after orientation change Add missing nullify for the s60 style's background pixmap pointer. Fixes QT-4753 sort out stdout vs. stderr usage Changes for 4.7.3 Implement tiled image and pixmap drawing in VG paint engine. Changed maximum heap size for qmlflickr on Symbian. Change Symbian to use destroyed swap behaviour in GL Remove useless profile reference.
| * \ \ Merge remote-tracking branch 'origin/4.7' into qt-master-from-4.7Thiago Macieira2011-03-2417-85/+283
| |\ \ \ |/ / / / | | | | | | | | | | | | Conflicts: src/opengl/qwindowsurface_gl.cpp
| * | | Merge branch '4.7' of scm.dev.nokia.troll.no:qt/qt-s60-public into ↵Qt Continuous Integration System2011-03-233-12/+40
| |\ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 4.7-integration * '4.7' of scm.dev.nokia.troll.no:qt/qt-s60-public: Icon in QMessageBox::question is deformed on Symbian Prevent crash when connecting signal in splitview GraphicsViews: Focused widget is not visible after orientation change GraphicsViews: Focused widget is not visible after orientation change Add missing nullify for the s60 style's background pixmap pointer.
| | * | | Icon in QMessageBox::question is deformed on SymbianSami Merila2011-03-231-3/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | MessageBox icons are deformed on Qt/S60, since style creates all icons with 1:1 width:height ratio. Based on S60 documentation messagebox icons are of 4:3 ratio. This is the same ratio between all the S60 releases. Task-number: QTBUG-14953 Reviewed-by: Laszlo Agocs
| | * | | Prevent crash when connecting signal in splitviewSami Merila2011-03-231-9/+12
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If there is no focusitem when connecting a signal, splitview crashes. This is caused by null pointer use. Prevent the crash by checking the pointer. Task-number: QTBUG-16572 Reviewed-by: Guoqing Zhang
| | * | | GraphicsViews: Focused widget is not visible after orientation changeSami Merila2011-03-231-12/+14
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Clean up previous fix a little bit, by merging if-statements. Task-number: QTBUG-17576 Reviewed-by: Guoqing Zhang
| | * | | GraphicsViews: Focused widget is not visible after orientation changeSami Merila2011-03-231-0/+15
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If there is a graphicsview with input widget focused and splitview is open, and we rotate the device from Portrait to Landscape (or vice versa), the input widget is no longer visible in the splitview rect. This only happens if there are scrollbars visible in the view. Reason is that the input widget is tried to ensure visibility for before window surface is valid. In these cases, make the input widget visibility re-location happen after window server indicates that the view is now visible. Task-number: QTBUG-17576 Reviewed-by: Guoqing Zhang
| | * | | Add missing nullify for the s60 style's background pixmap pointer.Laszlo Agocs2011-03-231-0/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Qt apps would crash during orientation change without setting m_background to null after deleting it. Reviewed-by: Sami Merila
| * | | | Merge branch '4.7' of scm.dev.nokia.troll.no:qt/oslo-staging-1 into ↵Qt Continuous Integration System2011-03-232-10/+13
| |\ \ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 4.7-integration * '4.7' of scm.dev.nokia.troll.no:qt/oslo-staging-1: Fixes QT-4753
| | * | | | Fixes QT-4753Xizhi Zhu2011-03-232-10/+13
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Invalid read in ICD plugin for bearer management. Reviewed-by: Iiro Kause
| * | | | | Merge branch '4.7' of scm.dev.nokia.troll.no:qt/oslo-staging-1 into ↵Qt Continuous Integration System2011-03-223-34/+86
| |\ \ \ \ \ | | |/ / / / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 4.7-integration * '4.7' of scm.dev.nokia.troll.no:qt/oslo-staging-1: sort out stdout vs. stderr usage Changes for 4.7.3
| | * | | | sort out stdout vs. stderr usageOswald Buddenhagen2011-03-222-34/+39
| | | | | | | | | | | | | | | | | | | | | | | | Task-number: QTBUG-14728
| | * | | | Changes for 4.7.3Timo Turunen2011-03-221-0/+47
| | | | | | | | | | | | | | | | | | | | | | | | Reviewed-by: Trust Me
| * | | | | Merge branch '4.7' of scm.dev.nokia.troll.no:qt/qt-s60-public into ↵Qt Continuous Integration System2011-03-229-31/+151
| |\ \ \ \ \ | | | |/ / / | | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 4.7-integration * '4.7' of scm.dev.nokia.troll.no:qt/qt-s60-public: Implement tiled image and pixmap drawing in VG paint engine. Changed maximum heap size for qmlflickr on Symbian. Change Symbian to use destroyed swap behaviour in GL Remove useless profile reference.
| | * | | | Implement tiled image and pixmap drawing in VG paint engine.Laszlo Agocs2011-03-221-3/+98
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The vgWritePixel-based "fallback" in drawImage can only be used when the transformation is simple and no opacity is set. The tiled drawing introduced here will serve as an additional fallback for the more complex cases and will allow drawing large (12 Mpix or more) pixmaps and images with transformation and semi-transparency even when the complete image data would not fit into the GPU memory. Task-number: QTBUG-18251 Reviewed-by: Jani Hautakangas
| | * | | | Merge branch '4.7' of git://scm.dev.nokia.troll.no/qt/qt-s60-public into 4.7Laszlo Agocs2011-03-227-27/+51
| | |\ \ \ \
| | | * | | | Change Symbian to use destroyed swap behaviour in GLJani Hautakangas2011-03-225-23/+51
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Destroyed swap behaviour is a little bit faster than preserved swap with heavily animated UIs like QML. Task-number: QTBUG-18270 Reviewed-by: Samuel Rødal
| | | * | | | Remove useless profile reference.axis2011-03-222-4/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The profile has been empty since the plugin functionality was moved into QtMultimedia itself (d73df73c556b0a225fdde7341549ed834bc691f4). RevBy: Miikka Heikkinen
| | * | | | | Changed maximum heap size for qmlflickr on Symbian.Laszlo Agocs2011-03-221-1/+2
| | |/ / / / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 32 MB is certainly not enough to load large (e.g. 12 Mpix or bigger) images. Now it is changed to 128 MB. Reviewed-by: Jani Hautakangas
* | | | | | Merge branch 'master' of scm.dev.nokia.troll.no:qt/oslo-staging-1 into ↵Qt Continuous Integration System2011-03-2451-37/+73633
|\ \ \ \ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | master-integration * 'master' of scm.dev.nokia.troll.no:qt/oslo-staging-1: (48 commits) Remove the use of the QtTest baseline feature I reverted Revert "Add a feature to QTestLib to correct benchmark results." Use the ARM version of UTF-8 detection in the Neon code Update the data files for the QString benchmark test Fix a bug in the SSE2 UTF-8 decoder. Add ARM Neon versions of fromLatin1 and fromUtf8 Make this compile on Atom/Core2 (no SSE4) and on ARM (no SSE) Add baselines and zeros to the benchmarks. Add an UTF-8 conversion on trusted data and no BOM. Make it easier to write a UTF-8 conversion on trusted data Add the missing tests and 4-byte UTF-8 sequences Improve the code and avoid unnecessary stores Add an UTF-8 conversion optimised for ASCII using SSE2 Add an UTF-8 conversion code that is optimised for ASCII Add a stateless copy of the Qt 4.7 UTF-8 codec. Add UTF-8 code benchmarks Improve a little more the core loop and propagate to the other code Reduce the number of operations in the main loop Add an SSE4 version using PMOVZXBW and PSRLDQ Attempt to improve the epilog code ...
| * | | | | | Remove the use of the QtTest baseline feature I revertedThiago Macieira2011-03-221-4/+4
| | | | | | |
| * | | | | | Revert "Add a feature to QTestLib to correct benchmark results."Thiago Macieira2011-03-2220-134/+33
| | | | | | | | | | | | | | | | | | | | | | | | | | | | This reverts commit 3cc83349d9a07bcb7ae9c338bfdc355905a5a2f1.
| * | | | | | Use the ARM version of UTF-8 detection in the Neon codeThiago Macieira2011-03-221-16/+13
| | | | | | |
| * | | | | | Update the data files for the QString benchmark testThiago Macieira2011-03-223-204/+71503
| | | | | | | | | | | | | | | | | | | | | | | | | | | | Also add the Neon build flags
| * | | | | | Fix a bug in the SSE2 UTF-8 decoder.Thiago Macieira2011-03-221-0/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If the multibyte sequence crossed the end of the SSE2 area, then len would have a wrong value and the data wouldn't be read.
| * | | | | | Add ARM Neon versions of fromLatin1 and fromUtf8Thiago Macieira2011-03-221-0/+192
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The fromLatin1 code is very simple, yet the handwritten assembly performs better due to the use of post-increments. The fromUtf8 code has two alternatives. Neon lacks an instruction similar to SSE2's _mm_movemask_epi8 (PMOVMSKB) which extracts one bit from each byte and stores it in a register. We used that in the UTF-8 code to detect bytes with the highest bit set. To compensate, we used two alternatives: 1) AND the comparison result with a vector containing {128, 64, ...,1 } Do 3 parallel-adds (VPADD.I8), which will make the mask propagate to the lowest component in the vector. Trick found in: http://hilbert-space.de/?p=22 (comment 16-17) 2) Extract the two words from the doubleword Neon register and do the work in ARM assembly. It looks like the latter version is performing better.
| * | | | | | Make this compile on Atom/Core2 (no SSE4) and on ARM (no SSE)Thiago Macieira2011-03-221-1/+18
| | | | | | |
| * | | | | | Add baselines and zeros to the benchmarks.Thiago Macieira2011-03-221-6/+13
| | | | | | |
| * | | | | | Add an UTF-8 conversion on trusted data and no BOM.Thiago Macieira2011-03-221-0/+48
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This assumes that there are no overlong sequences, no continuation characters without the leading, no missing continuations and no BOM.
| * | | | | | Make it easier to write a UTF-8 conversion on trusted dataThiago Macieira2011-03-221-38/+58
| | | | | | |
| * | | | | | Add the missing tests and 4-byte UTF-8 sequencesThiago Macieira2011-03-221-1/+23
| | | | | | |
| * | | | | | Improve the code and avoid unnecessary storesThiago Macieira2011-03-221-7/+9
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If there's an UTF-8 high byte in the first 8 bytes, don't try to save the latter 8 characters
| * | | | | | Add an UTF-8 conversion optimised for ASCII using SSE2Thiago Macieira2011-03-221-1/+54
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This code is 2x faster than the original UTF-8 code and within 35% of the pure Latin1 code
generated by cgit v0.12 at 2024-12-28 22:37:16 (GMT)