summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Improved server logging and more robust client connection attemptaavit2010-09-095-27/+38
| | | | Also misc minor improvements all around
* Opened for running on Mac tooaavit2010-09-081-2/+5
|
* Merge branch 'lancelot' of scm.dev.nokia.troll.no:qt/aavits-graphicsstuff ↵aavit2010-09-081-1/+3
|\ | | | | | | into lancelot
| * Check the QT_LANCELOT_SERVER env var for the servername.Trond Kjernåsen2010-09-081-1/+3
| |
* | Shave 25% off test time by smarter qps script parsingaavit2010-09-082-3/+26
|/
* Fix: remove baseline checksums before adding renderedaavit2010-09-083-26/+1
| | | | | | .. otherwise, the comparison will always yield true... Also a simplification: QList provides default datastreaming ops
* It's the GL_VERSION string that usually contains the Mesa substring.Trond Kjernåsen2010-09-071-1/+1
| | | | | | This of course differs, depending on the different backends used, but so far it seems most of the Mesa implementations have Mesa mentioned in the version string.
* Make it possible to have several checksums associated with one ImageItem.Trond Kjernåsen2010-09-074-20/+44
| | | | | Also changed the meta file from being a text file to a binary file. Nothing uses the extra checksums yet, but that should be added soon.
* Add another format, and blacklist some moreaavit2010-09-072-1/+22
|
* More server storage path fixingaavit2010-09-072-7/+16
|
* Improve servers storage directory structureaavit2010-09-073-10/+38
| | | | | - support different pixel formats for same engine - separate baselines per hostname
* Avoid testing on Mesaaavit2010-09-071-34/+48
|
* Merge branch 'lancelot' of scm.dev.nokia.troll.no:qt/aavits-graphicsstuff ↵aavit2010-09-071-1/+8
|\ | | | | | | into lancelot
| * We only want to test hw accelerated GL 2.0 implementations.Trond Kjernåsen2010-09-061-1/+8
| |
* | Minor style improvementsaavit2010-09-071-8/+6
|/
* Merge branch 'lancelot' of scm.dev.nokia.troll.no:qt/aavits-graphicsstuff ↵aavit2010-09-033-16/+17
|\ | | | | | | into lancelot
| * engineAsString() should've been const.Trond Kjernåsen2010-09-032-2/+2
| |
| * Move the engine->string function into ImageItem.Trond Kjernåsen2010-09-033-16/+17
| |
* | Proper checksum function implementedaavit2010-09-033-3/+777
|/
* Data column names have apparently changed :)Trond Kjernåsen2010-09-031-1/+1
|
* Add basic OpenGL testing support.Trond Kjernåsen2010-09-033-10/+83
|
* Protocol rewrite done; design cleaned up.aavit2010-09-035-242/+95
| | | | | | Client no longer fetches baseline images, not even on mismatch. Bitwise comparison, mismatch score etc. will be done on server, if wanted.
* Rewrite protocol (ImageItem/ImageItemList)aavit2010-09-025-90/+265
| | | | Not finished, but core functionality is in working state.
* try a different image formataavit2010-08-311-3/+3
|
* Add reporting of difference score on mismatchesaavit2010-08-301-15/+57
|
* Experiment with generating and storing the diff image also on failureaavit2010-08-301-0/+13
|
* minor message fixaavit2010-08-271-1/+1
|
* Use a remote test server during dev, plus some housekeepingaavit2010-08-256-12/+22
|
* First shot at a new client-server based autotest of the Arthuraavit2010-08-2570-0/+7741
| | | | | 2D rendering system, based on the lance test tool. Hence dubbed lancelot.
* Fix tabArray support for boundingRect measurementJiang Jiang2010-08-231-1/+8
| | | | | | | | | | | QFontMetrics::boundingRect() and size() accept a tabArray as argument to measure the size of string, but tabArray argument has no effect because qt_format_text() just ignore that. This patch make it handle tabArray so that measurement for tab aligned text can be handled correctly. Task-number: QTBUG-4904 Reviewed-by: Eskil
* Make the QItemSelectionRange and QItemSelectionModel ensure that the model ↵Stephen Kelly2010-08-192-1/+30
| | | | | | | is correct. Merge-request: 720 Reviewed-by: Gabriel de Dietrich <gabriel.dietrich-de@nokia.com>
* Fixed autotest compilation issue on solarisThierry Bastian2010-08-195-19/+19
| | | | Task-number: QTBUG-12991
* Fixed build issue on SolarisThierry Bastian2010-08-191-3/+3
| | | | Task-number: QTBUG-12994
* Merge branch 'master' of scm.dev.nokia.troll.no:qt/oslo-staging-2 into ↵Qt Continuous Integration System2010-08-191-0/+7
|\ | | | | | | | | | | | | master-integration * 'master' of scm.dev.nokia.troll.no:qt/oslo-staging-2: QGroupBox: Fix the focus rect of the title when using stylesheet
| * QGroupBox: Fix the focus rect of the title when using stylesheetMartin Pejcoch2010-08-191-0/+7
| | | | | | | | | | Task-number: QTBUG-13015 Reviewed-by: Olivier Goffart
* | Merge branch 'master' of scm.dev.nokia.troll.no:qt/oslo-staging-1 into ↵Qt Continuous Integration System2010-08-1915-46/+60
|\ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | master-integration * 'master' of scm.dev.nokia.troll.no:qt/oslo-staging-1: qUncompress: don't exit on allocation failure with -no-exceptions Prefer Q_CHECK_PTR to q_check_ptr Prefer Q_CHECK_PTR to q_check_ptr Ensure Q_CHECK_PTR always requires terminating ; Q_CHECK_PTR should issue qFatal with QT_NO_EXCEPTIONS Workaround what seems to be a compiler parser bug
| * | qUncompress: don't exit on allocation failure with -no-exceptionsJoão Abecasis2010-08-192-26/+22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If we're not allowed to fail here with -no-exceptions, we shouldn't throw either so just check the return value from qRealloc and be done with it. Also changed the test case so it doesn't expect bad_alloc being thrown; the test data now gets properly initialized... Reviewed-by: Olivier Goffart Reviewed-by: Harald Fernengel
| * | Prefer Q_CHECK_PTR to q_check_ptrJoão Abecasis2010-08-1910-16/+32
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | With the macro version we get proper debug output (with __FILE__, __LINE__ annotations) when compiling with -no-exceptions and !QT_NO_DEBUG. All changes in this patch affect assignment to local variables, where the order of assignment versus actual pointer check is not important. Reviewed-by: Olivier Goffart
| * | Prefer Q_CHECK_PTR to q_check_ptrJoão Abecasis2010-08-191-1/+3
| | | | | | | | | | | | | | | | | | | | | | | | With the macro version we get proper debug output (with __FILE__, __LINE__ annotations) when compiling with -no-exceptions and !QT_NO_DEBUG. Reviewed-by: Olivier Goffart
| * | Ensure Q_CHECK_PTR always requires terminating ;João Abecasis2010-08-191-1/+1
| | | | | | | | | | | | | | | | | | ... even with -no-exceptions and QT_NO_DEBUG. Reviewed-by: Olivier Goffart
| * | Merge branch 'master' of scm.dev.nokia.troll.no:qt/oslo-staging-1 into ↵Qt Continuous Integration System2010-08-192-2/+2
| |\ \ |/ / / | | | | | | | | | | | | | | | | | | master-integration * 'master' of scm.dev.nokia.troll.no:qt/oslo-staging-1: Q_CHECK_PTR should issue qFatal with QT_NO_EXCEPTIONS Workaround what seems to be a compiler parser bug
| * | Q_CHECK_PTR should issue qFatal with QT_NO_EXCEPTIONSJoão Abecasis2010-08-171-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Allowing the null pointer to propagate seems to be a bigger evil and could potentially become a security risk. Task-number: QT-3785 Reviewed-by: Peter Hartmann Reviewed-by: Olivier Goffart Reviewed-by: Thiago Macieira
| * | Workaround what seems to be a compiler parser bugJoão Abecasis2010-08-171-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The workaround seems harmless, otherwise. Maybe even standard-compliant (14.2.4 [temp.names]) $ gcc --version sbox-arm-none-linux-gnueabi-gcc (cs2009q1-eglibc2.8-armv7) 4.3.3 Copyright (C) 2008 Free Software Foundation, Inc. Reviewed-by: Olivier Goffart
* | | Merge branch 'master' of scm.dev.nokia.troll.no:qt/oslo-staging-2 into ↵Qt Continuous Integration System2010-08-1913-76/+80
|\ \ \ | | |/ | |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | master-integration * 'master' of scm.dev.nokia.troll.no:qt/oslo-staging-2: Disable translatedRotatedAndScaledPainter testcase as it is broken. Fixed partial update on QLineEdit disable the outlinefillconsistency test until the line algorithm has been updated. Fix nasty copy-paste bug in fetchTransformedBilinear() fix breakages in qpainter autotests. QMetaType: add "signed char" as builtin. Fix how subpixel positions are intepreted in an aliased grid.
| * | Merge branch 'master' of git@scm.dev.nokia.troll.no:qt/oslo-staging-2Gunnar Sletta2010-08-193-8/+19
| |\ \
| | * | Fixed partial update on QLineEditThierry Bastian2010-08-183-8/+19
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | QLineControl passes a QRect to tell when a partial update is needed. We used to connect this slot to update. This made the whole lineedit update when the cursor blinked. Task-number: QTBUG-12709 Reviewed-by: gabi
| * | | Disable translatedRotatedAndScaledPainter testcase as it is broken.Gunnar Sletta2010-08-191-1/+20
| |/ / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The problem is that the order of rounding to fixedpoint is different in the static and normal text case and this leads to some glyphs occationally being 1 pixel off. In the normal case, the glyphs are laid out and positioned in fixed point, then transformed using the full current matrix including the position which is then snapped to integer device space when drawn. In the static text case, the glyphs are laid out and positioned in fixed point related to 0,0. The actual device position is then added to the glyph positions.
| * | Merge branch 'master' of scm.dev.nokia.troll.no:qt/oslo-staging-2Gunnar Sletta2010-08-186-29/+180
| |\ \
| | * | Fix nasty copy-paste bug in fetchTransformedBilinear()Paul Olav Tvete2010-08-171-2/+2
| | | | | | | | | | | | | | | | | | | | | | | | Introduced in commit 0d7e68391 Reviewed-by: Samuel
| | * | Merge branch 'master' of scm.dev.nokia.troll.no:qt/oslo-staging-2 into ↵Qt Continuous Integration System2010-08-178-65/+36
| | |\ \ | |/ / / |/| | | | | | | | | | | | | | | | | | | | | | | | | | | master-integration * 'master' of scm.dev.nokia.troll.no:qt/oslo-staging-2: fix breakages in qpainter autotests. QMetaType: add "signed char" as builtin. Fix how subpixel positions are intepreted in an aliased grid.
generated by cgit v0.12 at 2025-09-07 18:31:07 (GMT)