summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* prefer DerivedNormalizationProps.txt over CompositionExclusions.txtRitt Konstantin2010-03-052-208/+28
| | | | | | | | | | | | | | | | | | | 1) http://www.unicode.org/reports/tr44/ :2.1 > Implementations should simply use the derived properties, > and should not try to rederive them from lists of simple > properties and collections of rules, because of the chances > for error and divergence when doing so. 2) DerivedNormalizationProps.txt file also provides additional info that can be used for Normalization Form Quick Check algorithm implementation some later Note: this commit *must not* change anything in the (re)generated data; any change in generated data will point to the data inconsistency between CompositionExclusions.txt, DerivedNormalizationProps.txt and UnicodeData.txt files Merge-request: 480 Reviewed-by: Thiago Macieira <thiago.macieira@nokia.com>
* improve error reporting a bit moreRitt Konstantin2010-03-051-1/+8
| | | | | Merge-request: 480 Reviewed-by: Thiago Macieira <thiago.macieira@nokia.com>
* improve error reportingRitt Konstantin2010-03-051-0/+8
| | | | | | | | warn on upperCaseDiff/lowerCaseDiff/titleCaseDiff/caseFoldDiff 's value exceeds signed short bounds Merge-request: 480 Reviewed-by: Thiago Macieira <thiago.macieira@nokia.com>
* fix incorect conditionRitt Konstantin2010-03-051-1/+1
| | | | | Merge-request: 480 Reviewed-by: Thiago Macieira <thiago.macieira@nokia.com>
* check if string to int conversions were done w/o errorsRitt Konstantin2010-03-051-3/+18
| | | | | Merge-request: 480 Reviewed-by: Thiago Macieira <thiago.macieira@nokia.com>
* improve error reporting for unassigned grapheme/word/sentence break classesRitt Konstantin2010-03-051-8/+18
| | | | | Merge-request: 480 Reviewed-by: Thiago Macieira <thiago.macieira@nokia.com>
* avoid using of qunicodetables_p.h in generatorRitt Konstantin2010-03-051-1/+3
| | | | | | | we can do that after last changes Merge-request: 480 Reviewed-by: Thiago Macieira <thiago.macieira@nokia.com>
* use QHash for line break mapRitt Konstantin2010-03-051-45/+74
| | | | | | | | to be consistent with other maps this also improves perfomance a bit Merge-request: 480 Reviewed-by: Thiago Macieira <thiago.macieira@nokia.com>
* use QHash for age mapRitt Konstantin2010-03-051-26/+29
| | | | | | | to be consistent with other maps Merge-request: 480 Reviewed-by: Thiago Macieira <thiago.macieira@nokia.com>
* minor clean-upsRitt Konstantin2010-03-051-42/+44
| | | | | | | | | | no changes in behavior in our case `c->cat != QChar::NoCategory` logically equals to `c->name` and the latter expression used in other maps Merge-request: 480 Reviewed-by: Thiago Macieira <thiago.macieira@nokia.com>
* generate specialCaseMap as set of human-readable code-sequencesRitt Konstantin2010-03-051-3/+3
| | | | | Merge-request: 480 Reviewed-by: Thiago Macieira <thiago.macieira@nokia.com>
* make sure that the sequences in specialCaseMap are uniqueRitt Konstantin2010-03-051-1/+1
| | | | | | | | | there is was a bug caused appending the same sequence again and again... this fix also reduces count of unique unicode properties (and blocks) Merge-request: 480 Reviewed-by: Thiago Macieira <thiago.macieira@nokia.com>
* Unicode character property 'General_Category=Cn' was erroneously ignoredRitt Konstantin2010-03-052-1/+14
| | | | | | | | | | | | | causing two related bugs in QChar::category() and QChar::isPrint(). As described in tr44, 4.2.8 Default Values General_Category character property should be setted to Cn for all codepoints omitted in UCD. Instead they was mapped to QChar::NoCategory (=0) which have no equivalent in Unicode specs and as result was ignored in filtering by binary mask. Merge-request: 480 Reviewed-by: Thiago Macieira <thiago.macieira@nokia.com>
* fix some 0x10000 codepoint -related issuesRitt Konstantin2010-03-051-10/+8
| | | | | Merge-request: 480 Reviewed-by: Thiago Macieira <thiago.macieira@nokia.com>
* minor cleanups and styling fixes for generator and generated codeRitt Konstantin2010-03-051-92/+98
| | | | | Merge-request: 480 Reviewed-by: Thiago Macieira <thiago.macieira@nokia.com>
* apply some (forgotten) changes to generator as they were applied to the dataRitt Konstantin2010-03-051-3/+9
| | | | | | | | b6542eb2 - Thierry Bastian) fix to exported symbol 2e429e40 - Lars Knoll) N'Ko support Merge-request: 480 Reviewed-by: Thiago Macieira <thiago.macieira@nokia.com>
* adjust unicode tables generator to work better under winRitt Konstantin2010-03-052-0/+2
| | | | | | | | | * protect data-files from crlf line-endings since generator asserts on windows because of crlf line endings * assume generator is a console app Merge-request: 480 Reviewed-by: Thiago Macieira <thiago.macieira@nokia.com>
* Make the UTF-8 encoder/decoder not accept Unicode non-charactersThiago Macieira2010-03-052-16/+97
| | | | Reviewed-By: Denis Dzyubenko
* Finish reverting the UTF-8 hack added in 80ea01c6 (P4 106704, Qt 3.2).Thiago Macieira2010-03-051-9/+2
| | | | | | | This code was removed from QString in 539cd1e5 (P4 259474, Qt 4.3), but apparently lingered on the UTF-8 codec code. Reviewed-by: Denis Dzyubenko
* Make QString::toUtf8() also use QUtf8.Thiago Macieira2010-03-051-39/+5
| | | | Reviewed-By: Denis Dzyubenko
* Doc: add some notes about QString lossy/lossless conversionsThiago Macieira2010-03-051-5/+38
|
* Doc: fix a mistake saying that const char* was converted to QLatin1StringThiago Macieira2010-03-051-5/+5
|
* Doc: Update QString docs about real ASCII and to/fromAscii and NULs.Thiago Macieira2010-03-051-31/+47
| | | | | | | | | | | | | | | Despite the names, the QString::toAscii and QString::fromAscii functions are not about ASCII. One can set an ASCII-incompatible codec using QTextCodec::setCodecForCStrings, which may lead to loss of information. Also update the docs about how the functions that take QByteArray take embedded NULs in the QByteArray into account. The result is: stop conversion at NUL: operator=, operator==, operator!= include NUL in conversion: operator<, operator<=, operator>, operator>= Reviewed-by: ossi
* FreeType: Fast lookup of Latin-1 glyphsAndreas Kling2010-03-054-18/+72
| | | | | | Store all glyphs with index < 256 in a pointer array for fast lookup Reviewed-by: Benjamin Poulain
* Assistant: Fix warnings about unused variables.ck2010-03-053-4/+4
|
* Fixing compilation error, missing ifdef.Carlos Manuel Duclos Vergara2010-03-051-1/+2
| | | | Reviewed-by: Prasanth
* Mac: qfontdialog autotest failesRichard Moe Gustavsen2010-03-051-9/+33
| | | | | | | | | | The reason is the way we leave modality in the font dialog. When we do, we need to close the non-native qfontdialog as well, but doing so will issue callbacks (like set_visible_sys(false)), that will cause problems. This patch makes sure we wait with such callbacks until the native modal event loop has finished executing Reviewed-by: cduclos
* Mac: non-modal native qfontdialog does not work at allRichard Moe Gustavsen2010-03-054-54/+135
| | | | | | | | | | This was never implemented, and as it stands, just creating a QFontDialog and calling show on it will not show anything. This patch makes sure that show works, and that one can have more than one font dialog showing at the same time (but only the first one will be native) Reviewed-by: cduclos
* Mac: qfontdialog test failsRichard Moe Gustavsen2010-03-051-1/+1
| | | | | | | | The native font dialog on mac did not set the font specified in the non-native font dialog upon exec, but rather qApp->font(). This oneliner sets the correct font, and makes the autotest pass. Reviewed-by: cduclos
* Mac: QFontDialog will not sometimes leave modalityRichard Moe Gustavsen2010-03-051-26/+9
| | | | | | | | It turns out the the code implemented two different ways of entering modality, and mixed the two when it came to exit modality. This patch just uses the nsapp runmodalforwindow approach. Reviewed-by: cduclos
* Mac: static getFont functions in QFontDialog does not workRichard Moe Gustavsen2010-03-053-203/+0
| | | | | | | | | | | | This bug is visible if the static funtions are used before calling qApp->exec(). The reason is that the static functions still use the old code path, rather than the new one (that will end up creating a QFontDialog and call exec on it). Just using the new style will fix the problem. Task-number: QTBUG-7371 Reviewed-by: cduclos
* Merge branch '4.7' of git@scm.dev.nokia.troll.no:qt/oslo-staging-1 into 4.7Martin Smith2010-03-055-58/+82
|\
| * Don't needlessly call pushContext() when reading propertiesKent Hansen2010-03-051-15/+18
| | | | | | | | | | | | | | It's only necessary to push a QScriptContext when properties are read from an object that inherits QScriptable. Postpone the decision of pushing a context until we know whether the object is a QScriptable.
| * QtScript: Don't needlessly make deep copies of function namesKent Hansen2010-03-052-35/+55
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The QObject binding uses the function name to determine whether a meta-method is an overload. This was implemented quite naively by copying the name from the signature into another array before comparing it. This could cause several unnecessary memory allocations, since storing the name is really only needed when there is an exception (e.g. ambiguous call). Instead, use strncmp to compare only the relevant characters of the original (non-copied) method signature. This makes normal slot invocation from QtScript up to 15% faster. Reviewed-by: Jedrzej Nowacki
| * QMetaType::type(): return immediately if the typename is emptyKent Hansen2010-03-051-0/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | No need to look up / lock data structures if we know that the comparisons will all fail. This was encountered because QMetaMethod::typeName() returns an empty string if the return type is void (even though there is a QMetaType::Void (value 0) with name "void"). This was causing the QtScript meta-object binding to spend a lot of its time looking up the type for an empty string when invoking slots that return void. Rather than having these checks in QtScript and who knows where else, it's better that QMetaType::type() does it itself. No regressions in the qmetatype benchmark. Reviewed-by: Olivier Goffart Reviewed-by: Harald Fernengel
| * Change behavior of applicationShouldTerminate.Fabien Freling2010-03-051-8/+5
| | | | | | | | | | | | | | | | We now terminate the application as long as Qt says we can. If not, then we cancel the termination. Task-number: QTBUG-6296 Reviewed-by: Morten Sorvig
| * Prevent a freeze of QFileSystemWatcher on Mac.Denis Dzyubenko2010-03-051-0/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | On Mac when the FSEvents backend is used and a file is added or removed from a file system watcher, we need to wait until the thread is finished, otherwise it is possible that the thread already exited from the run() function but hasn't fully terminated, meaning when we restart the thread by calling start() it won't start because QThread thinks it's already running. A better fix might be to avoid stopping and starting threads - to just stop the FSEvents loop and notify the thread that a file set has changed. Task-number: QTBUG-8524 Reviewed-by: Morten
* | doc: Fixed some QML qdoc errors.Martin Smith2010-03-055-8/+23
|/
* Merge branch '4.7' of git@scm.dev.nokia.troll.no:qt/oslo-staging-1 into 4.7Martin Smith2010-03-0517-11/+8914
|\
| * Merge branch '4.7' of scm.dev.nokia.troll.no:qt/oslo-staging-1 into ↵Qt Continuous Integration System2010-03-0524-184/+541
| |\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 4.7-integration * '4.7' of scm.dev.nokia.troll.no:qt/oslo-staging-1: Update the Qt focus widget on responder change. Doc: fix spelling errors. add wifi power notifications. more Qt way for autorelease. Remove unnecessary qWarnings from bearer management code. Don't keep polling network sessions open indefinitely. Prefer ethernet over wlan. Updating polling mechanism for bearer management engines.
| | * Update the Qt focus widget on responder change.Morten Johan Sørvig2010-03-051-0/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Set focus widget when the view becomes the first responder. This matches resignFirstResponder, where we set the focus widget to 0. This fixes the QCombobox::autoCompletionCaseSensitivity test failure introduced in the Mac alien widgets change. Reviewed-by: Richard Moe Gustavsen
| | * Doc: fix spelling errors.Morten Johan Sørvig2010-03-051-2/+2
| | |
| * | Merge branch '4.7' of scm.dev.nokia.troll.no:qt/qt-s60-public into ↵Qt Continuous Integration System2010-03-0515-9/+8905
| |\ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | 4.7-integration * '4.7' of scm.dev.nokia.troll.no:qt/qt-s60-public: Change WINSCW include paths to use same rules as ARMCC Create 4.7 def files for Symbian
| | * | Change WINSCW include paths to use same rules as ARMCCShane Kearns2010-03-041-1/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The default rules applied by symbian cause problems with headers included from within other headers. This only affects WINSCW builds, so the headers always need to be specially fixed up to be compatible, which is wrong. Instead, use CXXFLAGS to change the include path rules to be the same as Reviewed-by: Iain
| | * | Merge branch '4.7' of scm.dev.nokia.troll.no:qt/qt-s60-public into ↵Qt Continuous Integration System2010-03-0414-8/+8899
| | |\ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 4.7-integration * '4.7' of scm.dev.nokia.troll.no:qt/qt-s60-public: Create 4.7 def files for Symbian
| | | * | Create 4.7 def files for SymbianShane Kearns2010-03-0414-8/+8899
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | New QtDeclarative module New APIs in existing modules Reviewed-by: Trust Me
* | | | | doc: Fixed some QML qdoc errors.Martin Smith2010-03-052-31/+70
| |_|_|/ |/| | |
* | | | add wifi power notifications. more Qt way for autorelease.Lorn Potter2010-03-052-33/+83
| | | | | | | | | | | | | | | | and change to better way of getting wifi interfaces.
* | | | Merge remote branch 'staging/4.7' into bearermanagement/pollingAaron McCarthy2010-03-05630-14628/+19326
|\ \ \ \ | |/ / /
| * | | Merge branch '4.7' of scm.dev.nokia.troll.no:qt/oslo-staging-1 into ↵Qt Continuous Integration System2010-03-05115-10800/+12888
| |\ \ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 4.7-integration * '4.7' of scm.dev.nokia.troll.no:qt/oslo-staging-1: (72 commits) Doc: Enabled the context menu in the spreadsheet demo. Updated URLs. Doc: Applied suggestion for an improvement to the foreach documentation. Doc: Added a note about the lack of standard file copy/paste on X11. Doc: Fixed the row number for a widget in two parts of the tutorial. Doc: Fixed typos. Reduce compiling time of QScriptValue autotest suite on Windows. Optimize QScriptValue autotest generator. Set database write behavior to synchronous=OFF and increase page cache. doc: Fixed some qdoc errors. Setting ImhHiddenText for NoEcho line edits is not 100% correct, but still way better than fully visible text. Allow building documentation without all of Qt Added a documentation for the new enum value in gesture api. Remove the OBJECTS_DIR variable assignment from some projets in Qt. Fix compile qmake/MinGw: Link statically for Qt Creator to be able to detect it. Enable two fast path for blend_tiled_rgb565 Avoid QString reallocation for smallcaps fonts in Itemizer::generate() Make QLabel::text a reloadable property remove non wifi interfaces from being handled. ...
generated by cgit v0.12 at 2024-11-07 15:37:08 (GMT)