qt5base-lts/dist/changes-5.0.0

576 lines
24 KiB
Plaintext
Raw Normal View History

Some of the changes listed in this file include issue tracking numbers
corresponding to tasks in the Qt Bug Tracker:
http://bugreports.qt-project.org/
Each of these identifiers can be entered in the bug tracker to obtain more
information about a particular change.
****************************************************************************
* Source incompatible changes *
****************************************************************************
- QSslCertificate::subjectInfo() and QSslCertificate::issuerInfo() now
return a QStringList instead of a QString
- QSslCertificate::isValid() has been deprecated. Originally it only checked
the certificate dates, but later checking for blacklisting was added. Now
there's a more specific QSslCertificate::isBlacklisted() method.
- Unite clipping support has been removed from QPainter. The alternative is
to unite QRegion's and using the result on QPainter.
- QLibrary::resolve() now returns a function pointer instead of a void
pointer.
- QSslCertificate::alternateSubjectNames() is deprecated (but can be enabled
via QT_DISABLE_DEPRECATED_BEFORE), use
QSslCertificate::subjectAlternativeNames() instead.
- QLibraryInfo::buildKey() has been removed. Likewise, the QT_BUILD_KEY
preprocessor #define has also been removed. The build-key is obsolete
and is no longer necessary.
- QCoreApplication::translate() will no longer return the source text when
the translation is empty. Use lrelease -removeidentical for optimization.
- Qt::escape() is deprecated (but can be enabled via
QT_DISABLE_DEPRECATED_BEFORE), use QString::toHtmlEscaped() instead.
- QBool is gone. QString::contains, QByteArray::contains, and QList::contains
used to return an internal QBool class so that the Qt3 code
"if (a.contains() == 2)" wouldn't compile anymore. Such code cannot exist
in Qt4, so these methods return a bool now. If your code used the undocumented
QBool, simply replace it with bool.
- qIsDetached<> has been removed without replacement.
- QMetaType:
* QMetaType::construct() has been renamed to QMetaType::create().
* QMetaType::unregisterType() has been removed.
- QTestLib:
testlib: Count passes, fails and skips consistently. For data-driven tests, testlib previously counted one fail or skip for each data row that failed or skipped, while it counted only one pass for a test function where all rows passed and counted no passes for a test function where some rows passed and some rows failed. A similar problem also existed for benchmark tests, which could run multiple iterations of the same test, with each fail and skip being counted but only a single pass being counted for the entire series of iterations. This commit makes testlib count one pass, fail or skip for each data row. Test functions that are not data-driven count one result for the test function, as before. Benchmark tests count one pass, fail or skip per iteration. A side-effect of this change is that the test output in plain text, xml and light xml formats now shows a result for every data row and benchmark iteration executed, allowing post-processors to correctly calculate the total number of tests executed. Previously, individual rows were not shown in the test output if they passed, making such calculations impossible. The only change to the xunitxml output format is to correct a bug where no test result was recorded for a test function if the last data row was skipped and all other rows passed -- in which case the overall result should be a pass. Note that there is also a pre-existing bug in the xunit logger, where no result is reported if all rows are skipped; that bug is unaffected by this commit. Task-number: QTBUG-21848 Task-number: QTBUG-22124 Change-Id: I7e17177e10d6e89e55b9684c159bd506f21d002b Reviewed-by: Ed Baak <ed.baak@nokia.com>
2012-02-02 07:08:44 +00:00
* The plain-text, xml and lightxml test output formats have been changed to
show a test result for every row of test data in data-driven tests. In
Qt4, only fails and skips were shown for individual data rows and passes
were not shown for individual data rows, preventing accurate calculation
of test run rates and pass rates.
* The QTRY_VERIFY and QTRY_COMPARE macros have been moved into QTestLib.
These macros formerly lived in tests/shared/util.h but are now provided
by including the <QtTest/QtTest> header. In addition,
QTRY_VERIFY_WITH_TIMEOUT and QTRY_COMPARE_WITH_TIMEOUT are provided,
allowing for specifying custom timeout values.
* The QTEST_NOOP_MAIN macro has been removed from the API. If a test is
known at compile-time to be inapplicable for a particular build it should
be omitted via .pro file logic, or the test should call QSKIP in the
initTestCase() method to skip the entire test and report a meaningful
explanation in the test log.
* The DEPENDS_ON macro has been removed from the API. This macro did nothing
and misled some users to believe that they could make test functions depend
on each other or impose an execution order on test functions.
* The QTest::qt_snprintf function has been removed from the API. This was an
internal testlib function that was exposed in the public API due to its use
in a public macro. Any calls to this function should be replaced by a call
to qsnprintf(), which comes from the <QtCore/QByteArray> header.
* The QTest::pixmapsAreEqual() function has been removed from the API.
Comparison of QPixmap objects should be done using QCOMPARE, which provides
more informative output in the event of a failure.
* The QSKIP macro no longer has the "mode" parameter, which caused problems
for calculating test metrics, as the SkipAll mode hid information about
what test data was skipped. Calling QSKIP in a test function now behaves
like SkipSingle -- skipping a non-data-driven test function or skipping
only the current data row of a data-driven test function. Every skipped
data row is now reported in the test log.
- The QSsl::TlsV1 enum value was renamed to QSsl::TlsV1_0 .
- QAccessible:
* Internal QAccessible::State enum value HasInvokeExtension removed
- QAccessibleInterface:
* The "child" integer parameters have been removed. This moves the api
to be closer to IAccessible2.
This means several functions lose their integer parameter:
text(Text t, int child) -> text(Text t), rect(int child) -> rect()
setText(Text t, int child, const QString &text) -> setText(Text t, const QString &text)
role(int child) -> role(), state(int child) -> state()
* parent() and child() was added in order to do hierarchical navigation.
* relations() was added as a replacement to relationTo()
* As a consequence of the above two points, navigate() was removed.
* Accessible-Action related functions have been removed. QAccessibleInterface
subclasses are expected to implement the QAccessibleActionInterface instead.
These functions have been removed:
QAccessibleInterface::userActionCount, QAccessibleInterface::actionText,
QAccessibleInterface::doAction
- QAccessibleEvent also loses the child parameter.
QAccessibleEvent(Type type, int child) -> QAccessibleEvent(Type type)
QAccessibleEvent::child() removed.
- QAccessibleActionInterface:
* Refactored to be based on action names. All functions have been changed from using
int parameters to strings.
- QSound has been moved from QtGui to QtMultimedia
- QTabletEvent::QTabletEvent does not take a hiResGlobalPos argument anymore,
as all coordinates are floating point based now.
Extend touch events. The capability flags indicate which information is valid in the touch points. Previously there was no way to tell if e.g. the value returned by pressure() is actually the value provided by the driver/device or it is just something bogus due to pressure not being supported. The points' flags return information about the individual touch points. One use case is to differentiate between touches made by finger and pen. Velocity, if available, is now also exposed. Each touch point can now contain an additional list of "raw" positions. These points are not reported individually but are taken into account in some way by the underlying device and drivers to generate the final, "accurate" touch point. In case the underlying drivers expose these additional positions, they are made available in the lists returned by the touch points' rawScreenPosition(). The raw positions are only available in screen coordinates to prevent wasting time with mapping from global positions in applications that do not use this data. Instead, apps can query the QWindow to which the touch event was sent via QTouchEvent::window() and can call mapFromGlobal() manually if they need local raw positions. The capability and device type information is now held in a new QTouchDevice class. Each touch event will contain only a pointer to one of the global QTouchDevice instances. On top of type and capability, the new class also contains a name which can be used to differentiate between multiple touch input devices (i.e. to tell from which one a given QTouchEvent originates from). The introduction of QTouchDevice has three implications: The QTouchEvent constructor and QWindowSystemInterface::handleTouchEvent need to be changed (to pass a QTouchDevice pointer instead of merely a device type value), and each platform or generic plug-in is now responsible for registering one or more devices using the new API QWindowSystemInterface::registerTouchDevice. Change-Id: Ic1468d3e43933d8b5691d75aa67c43e1bc7ffe3e Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
2011-11-27 15:42:23 +00:00
- QTouchEvent:
* The DeviceType enum and deviceType() have been deprecated due to
the introduction of QTouchDevice.
* The signature of the constructor has changed. It now takes a
QTouchDevice pointer instead of just a DeviceType value.
* TouchPointState no longer includes TouchPointStateMask and
TouchPointPrimary. QTouchEvent::TouchPoint::isPrimary() has
been removed.
Extend touch events. The capability flags indicate which information is valid in the touch points. Previously there was no way to tell if e.g. the value returned by pressure() is actually the value provided by the driver/device or it is just something bogus due to pressure not being supported. The points' flags return information about the individual touch points. One use case is to differentiate between touches made by finger and pen. Velocity, if available, is now also exposed. Each touch point can now contain an additional list of "raw" positions. These points are not reported individually but are taken into account in some way by the underlying device and drivers to generate the final, "accurate" touch point. In case the underlying drivers expose these additional positions, they are made available in the lists returned by the touch points' rawScreenPosition(). The raw positions are only available in screen coordinates to prevent wasting time with mapping from global positions in applications that do not use this data. Instead, apps can query the QWindow to which the touch event was sent via QTouchEvent::window() and can call mapFromGlobal() manually if they need local raw positions. The capability and device type information is now held in a new QTouchDevice class. Each touch event will contain only a pointer to one of the global QTouchDevice instances. On top of type and capability, the new class also contains a name which can be used to differentiate between multiple touch input devices (i.e. to tell from which one a given QTouchEvent originates from). The introduction of QTouchDevice has three implications: The QTouchEvent constructor and QWindowSystemInterface::handleTouchEvent need to be changed (to pass a QTouchDevice pointer instead of merely a device type value), and each platform or generic plug-in is now responsible for registering one or more devices using the new API QWindowSystemInterface::registerTouchDevice. Change-Id: Ic1468d3e43933d8b5691d75aa67c43e1bc7ffe3e Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
2011-11-27 15:42:23 +00:00
* QWidget *widget() has been removed and is replaced by QObject
*target() in order to avoid QWidget dependencies.
* QEvent::TouchCancel has been introduced. On systems where it makes
sense this event type can be used to differentiate between a
regular TouchEnd and abrupt touch sequence cancellations caused by
the compositor, for example when a system gesture gets recognized.
- QMetaType
* Q_DECLARE_METATYPE(Foo*) now requires that Foo is fully defined. In
cases where a forward declared type should be used as a metatype,
Q_DECLARE_OPAQUE_POINTER(Foo*) can be used to allow that.
- QItemEditorFactory
* The signature of the createEditor and valuePropertyName methods
have been changed to take arguments of type int instead of QVariant::Type.
Extend touch events. The capability flags indicate which information is valid in the touch points. Previously there was no way to tell if e.g. the value returned by pressure() is actually the value provided by the driver/device or it is just something bogus due to pressure not being supported. The points' flags return information about the individual touch points. One use case is to differentiate between touches made by finger and pen. Velocity, if available, is now also exposed. Each touch point can now contain an additional list of "raw" positions. These points are not reported individually but are taken into account in some way by the underlying device and drivers to generate the final, "accurate" touch point. In case the underlying drivers expose these additional positions, they are made available in the lists returned by the touch points' rawScreenPosition(). The raw positions are only available in screen coordinates to prevent wasting time with mapping from global positions in applications that do not use this data. Instead, apps can query the QWindow to which the touch event was sent via QTouchEvent::window() and can call mapFromGlobal() manually if they need local raw positions. The capability and device type information is now held in a new QTouchDevice class. Each touch event will contain only a pointer to one of the global QTouchDevice instances. On top of type and capability, the new class also contains a name which can be used to differentiate between multiple touch input devices (i.e. to tell from which one a given QTouchEvent originates from). The introduction of QTouchDevice has three implications: The QTouchEvent constructor and QWindowSystemInterface::handleTouchEvent need to be changed (to pass a QTouchDevice pointer instead of merely a device type value), and each platform or generic plug-in is now responsible for registering one or more devices using the new API QWindowSystemInterface::registerTouchDevice. Change-Id: Ic1468d3e43933d8b5691d75aa67c43e1bc7ffe3e Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
2011-11-27 15:42:23 +00:00
- QWindowSystemInterface:
* The signature of all handleTouchEvent() variants have changed,
taking a QTouchDevice* instead of just a DeviceType value.
Platform or generic plug-ins have to create and register at least
one QTouchDevice before sending the first touch event.
* The event type parameter is removed from handleTouchEvent().
- The previously exported function qt_translateRawTouchEvent() has been removed.
Use QWindowSystemInterface::handleTouchEvent() instead.
- QAbstractEventDispatcher
* The signature for the pure-virtual registerTimer() has changed. Subclasses
of QAbstractEventDispatcher will need to be updated to reimplement the new
pure-virtual 'virtual void registerTimer(int timerId, int interval,
Qt::TimerType timerType, QObject *object) = 0;'
* QAbstractEventDispatcher::TimerInfo is no longer a QPair<int, int>. It is
now a struct with 3 members: struct TimerInfo { int timerId; int interval;
Qt::TimerType timerType; }; Reimplementations of
QAbstractEventDispatcher::registeredTimers() will need to be updated to pass
3 arguments to the TimerInfo constructor (instead of 2).
- QUuid
* Removed implicit conversion operator QUuid::operator QString(), instead
QUuid::toString() function should be used.
- The QHttp, QHttpHeader, QHttpResponseHeader and QHttpRequestHeader classes have
been removed, QNetworkAccessManager should be used instead.
- The QFtp class is no longer exported, QNetworkAccessManager should be used
instead.
- QProcess
* On Windows, QProcess::ForwardedChannels will not forward the output of GUI
applications anymore, if they do not create a console.
- QAbstractSocket's connectToHost() and disconnectFromHost() are now virtual and
connectToHostImplementation() and disconnectFromHostImplementation() don't exist.
- QTcpServer::incomingConnection() now takes a qintptr instead of an int.
- QNetworkConfiguration::bearerName() removed, and bearerTypeName() should be used.
- QDir::convertSeparators() (deprecated since Qt 4.2) has been removed. Use
QDir::toNativeSeparators() instead.
- QIconEngineV2 was merged into QIconEngine
You might need to adjust your code if it used a QIconEngine.
- qmake
* Projects which explicitly set an empty TARGET are considered broken now.
* The makespec and .qmake.cache do not see build pass specific variables any more.
* load()/include() with a target namespace and infile()/$$fromfile() now start with
an entirely pristine context.
* Configure's -sysroot and -hostprefix are now handled slightly differently.
The QT_INSTALL_... properties are now automatically prefixed with the sysroot;
the raw values are available as QT_INSTALL_.../raw and the sysroot as QT_SYSROOT.
The new QT_HOST_... properties can be used to refer to the Qt host tools.
* Several functions and built-in variables were modified to return normalized paths.
- QTextCodecPlugin has been removed since it is no longer used. All text codecs
are now built into QtCore.
- QDir::NoDotAndDotDot is QDir::NoDot|QDir::NoDotDot therefore there is no need
to use or check both.
Drop file-engine abstraction from public API This abstraction imposed serious performance penalties and is being dropped from the public API. In particular, by allowing file names to be arbitrarily hijacked by different file engines, and requiring engines to be instantiated in order to decide, it imposed unnecessary overhead on all file operations. Another flaw in the design with direct impact on performance is how engines have no way to provide (or retain) additional information obtained when querying the filesystem. In many places this has meant repeated operations on the file system, where useful information is immediately discarded to be queried again subsequently. For Qt 4.8 a major refactoring of the code base took place to allow bypassing the file-engine abstraction in select places, with considerable performance gains observed. In Qt 5 it is expected we'll be able to take this further, reaping even more benefits, but the abstraction has to go. [Dropping this now does not preclude that virtual file systems make an appearance in Qt at a later point in Qt 5's lifecycle. Hopefully with a new and improved abstraction.] Forward declarations for QFileExtension(Result) were dropped, as the classes were never used or defined. Tests using "internalized" classes will only fully run on developer builds. QFSFileEngine was removed altogether from exception safety test, as it isn't its intent to test internal API. Change-Id: Ie910e6c2628be202ea9e05366b091d6d529b246b Reviewed-by: Thiago Macieira <thiago.macieira@intel.com> Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
2012-02-17 19:09:17 +00:00
- QFSFileEngine, QAbstractFileEngine, QAbstractFileEngineIterator and
QAbstractFileEngineHandler were removed from public API and are no longer
exported. They may temporarily live as private implementation details, but
they may be altogether dropped or otherwise changed at will in the future.
- QLocale
* toShort(), toUShort(), toInt(), toUInt(), toLongLong() and toULongLong() no
longer take a parameter for base, they will only perform localised base 10
conversions. For converting other bases use the QString methods instead.
- QSystemLocale has been removed from the public API.
- QSqlQueryModel::indexInQuery() is now virtual. See note below under QSql.
- qMacVersion() has been removed. Use QSysInfo::macVersion() or
QSysInfo::MacintoshVersion instead.
****************************************************************************
* General *
****************************************************************************
General Improvements
--------------------
- The directory structure of the qtbase unit-tests has been reworked to
more closely match the directory structure of the code under test.
Integration tests have been moved to tests/auto/integrationtests.
Third party components
----------------------
-
****************************************************************************
* Library *
****************************************************************************
QtCore
------
* drop a bogus QChar::NoCategory enum value; the proper QChar::Other_NotAssigned
value is returned for an unassigned codepoints now.
* layoutAboutToBeChanged is no longer emitted by QAbstractItemModel::beginMoveRows.
layoutChanged is no longer emitted by QAbstractItemModel::endMoveRows. Proxy models
should now also connect to (and disconnect from) the rowsAboutToBeMoved and
rowsMoved signals.
* The default value of the property QSortFilterProxyModel::dynamicSortFilter was
changed from false to true.
* The signature of the virtual QAbstractItemView::dataChanged method has changed to
include the roles which have changed. The signature is consistent with the dataChanged
signal in the model.
* QFileSystemWatcher is now able to return failure in case of errors whilst
altering the watchlist in both the singular and QStringList overloads of
addPath and removePath.
* QString::mid, QString::midRef and QByteArray::mid, if the position passed
is equal to the length (that is, right after the last character/byte),
now return an empty QString, QStringRef or QByteArray respectively.
in Qt 4 they returned a null QString or a null QStringRef.
* QString methods toLongLong(), toULongLong(), toLong(), toULong(), toInt(),
toUInt(), toShort(), toUShort(), toDouble(), and toFloat() no longer use the
default or system locale, they will always use the C locale. This is to
guarantee consistent default conversion of strings. For locale-aware conversions
use the equivalent QLocale methods.
* QDate, QTime, and QDateTime have undergone important behavioural changes:
* QDate only implements the Gregorian calendar, the switch to the Julian
calendar before 1582 has been removed. This means all QDate methods will
return different results for dates prior to 15 October 1582, and there is
no longer a gap between 4 October 1582 and 15 October 1582.
* QDate::setYMD() is deprecated, use QDate::setDate() instead
* Most methods now apply strict validity checks and will return appropriate
and consistent values when invalid. For example, QDate::year() will return
0 and QDate::shortMonthName() will return QString().
* Adding days to a null QDate or seconds to a null QTime will no longer return
a valid QDate/QTime.
* QDate stores the Julian Day as a qint64 extending date support across a
more interesting range, see the class documentation for details.
* Conversion to YMD form dates is only accurate between to 4800 BCE to
1.4 million CE
* The QDate::addDays() and QDateTime::addDays() methods now take a qint64
* The QDate::daysTo() and QDateTime::daysTo() methods now return a qint64
* QTextCodec::codecForCStrings() and QTextCodec::setCodecForCStrings() have both
been removed. This was removed due to issues with breaking other code from
libraries, creating uncertainty/bugs in using QString easily, and (to a lesser
extent) performance issues.
* QTextCodec::codecForTr() and QTextCodec::setCodecForTr() have been removed,
QCoreApplication::Encoding value CodecForTr is now obsolete, use
DefaultCodec instead. For reasoning, see the codecForCStrings() removal above.
* QIntValidator and QDoubleValidator no longer fall back to using the C locale if
the requested locale fails to validate the input.
* A new set of classes for doing pattern matching with Perl-compatible regular
expressions has been added: QRegularExpression, QRegularExpressionMatch and
QRegularExpressionMatchIterator. They aim to replace QRegExp with a more
powerful and flexible regular expression engine.
* QEvent::AccessibilityPrepare, AccessibilityHelp and AccessibilityDescription removed:
* The enum values simply didn't make sense in the first place and should simply be dropped.
QtGui
-----
* Accessibility has been refactored. The hierachy of accessible objects is implemented via
proper parent/child functions instead of using navigate which has been deprecated for this purpose.
Table and cell interfaces have been added to qaccessible2.h
Extend touch events. The capability flags indicate which information is valid in the touch points. Previously there was no way to tell if e.g. the value returned by pressure() is actually the value provided by the driver/device or it is just something bogus due to pressure not being supported. The points' flags return information about the individual touch points. One use case is to differentiate between touches made by finger and pen. Velocity, if available, is now also exposed. Each touch point can now contain an additional list of "raw" positions. These points are not reported individually but are taken into account in some way by the underlying device and drivers to generate the final, "accurate" touch point. In case the underlying drivers expose these additional positions, they are made available in the lists returned by the touch points' rawScreenPosition(). The raw positions are only available in screen coordinates to prevent wasting time with mapping from global positions in applications that do not use this data. Instead, apps can query the QWindow to which the touch event was sent via QTouchEvent::window() and can call mapFromGlobal() manually if they need local raw positions. The capability and device type information is now held in a new QTouchDevice class. Each touch event will contain only a pointer to one of the global QTouchDevice instances. On top of type and capability, the new class also contains a name which can be used to differentiate between multiple touch input devices (i.e. to tell from which one a given QTouchEvent originates from). The introduction of QTouchDevice has three implications: The QTouchEvent constructor and QWindowSystemInterface::handleTouchEvent need to be changed (to pass a QTouchDevice pointer instead of merely a device type value), and each platform or generic plug-in is now responsible for registering one or more devices using the new API QWindowSystemInterface::registerTouchDevice. Change-Id: Ic1468d3e43933d8b5691d75aa67c43e1bc7ffe3e Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
2011-11-27 15:42:23 +00:00
* Touch events and points have been extended to hold additional
information like capability flags, point-specific flags, velocity,
and raw positions.
QtWidgets
---------
* QInputContext removed as well as related getters and setters on QWidget and QApplication.
Input contexts are now platform specific.
* QInputDialog::getInteger() has been obsoleted. Use QInputDialog::getInt() instead.
* In Qt 4, many QStyleOption subclasses were introduced in order to keep
binary compatibility -- QStyleOption was designed to be extended this way,
in fact it embeds a version number. In Qt 5 the various QStyleOption*V{2,3,4}
classes have been removed, and their members merged into the respective
base classes. Those classes were left as typedefs to keep existing code
working. Still, some minor adjustements could be necessary, especially in code
that uses QStyleOption directly and does not initialize all the members using
the proper Qt API: due to the version bump, QStyle will try to use the additional
QStyleOption members, which are left default-initialized.
* QHeaderView - The following functions have been obsoleted.
* void setMovable(bool movable) - use void setSectionsMovable(bool movable) instead.
* bool isMovable() const - use bool sectionsMovable() const instead.
* void setClickable(bool clickable) - use void setSectionsClickable(bool clickable) instead.
* bool isClickable() const - use bool sectionsClickable() instead.
* void setResizeMode(int logicalindex, ResizeMode mode) -
use setSectionResizeMode(logicalindex, mode) instead.
* ResizeMode resizeMode(int logicalindex) const -
use sectionResizeMode(int logicalindex) instead.
QtNetwork
---------
* QHostAddress::isLoopback() API added. Returns true if the address is
one of the IP loopback addresses.
* QSslCertificate::serialNumber() now always returns the serial number in
hexadecimal format.
* The openssl network backend now reads the ssl configuration file allowing
the use of openssl engines.
QtOpenGL
--------
QtScript
--------
QTestLib
--------
* [QTBUG-20615] Autotests can now log test output to multiple destinations
and log formats simultaneously.
QtSql
-----
QSqlTableModel/QSqlRelationalTableModel
* The dataChanged() signal is now emitted for changes made to an inserted
record that has not yet been committed. Previously, dataChanged() was
suppressed in this case for OnRowChange and OnFieldChange. This was probably
an attempt to avoid trouble if setData() was called while handling
primeInsert(). By emitting dataChanged(), we ensure that all views are aware
of the change.
* While handling primeInsert() signal, the record must be manipulated using
the provided reference. Do not attempt to manipulate the records using the
model methods setData() or setRecord().
* removeRows() no longer emits extra beforeDelete signal for out of range row.
* removeRows() now requires the whole range of targetted rows to be valid
before doing anything. Previously, it would remove what it could and
ignore the rest of the range.
* removeRows(), for OnFieldChange and OnRowChange, allows only 1 row to be
removed and only if there are no other changed rows.
* setRecord() and insertRecord()
-Only use fields where generated flag is set to true. This is
is consistent with the meaning of the flag.
-Require all fields to map correctly. Previously fields that didn't
map were simply ignored.
-For OnManualSubmit, insertRecord() no longer leaves behind an empty
row if setRecord() fails.
QSqlTableModel::setData()/setRecord(): fix incorrect row For OnFieldChange and OnRowChange, before submitting new changes, setData() and setRecord() attempt to submit pending changes and revert them upon failure. However, they fail to consider that reverting pending insertions removes rows from the model. As a result, the new change can be applied to a row higher than intended. One possible solution would be to adjust the targetted index for the removed rows, so that the intended row is affected by the new change. But this still causes the strange editing experience as rows jump up just as they are being edited. It does not seem right in the first place for the model to initiate reverting changes. It should be up to the application to decide what to do when data cannot be committed. In particular, setData() and setRecord() should not have the side effect of reverting already pending changes. The chosen solution is simply to refuse new changes that don't make sense for the edit strategy. For OnFieldChange, flag() will indicate read-only when editing is blocked by a pending change. Since setData() and setRecord() submit data immediately for OnFieldChange, it no longer makes sense to resubmit changes automatically before a new change. For OnRowChange, setData() keeps the behavior of automatically submitting a pending row before starting on a new row. This is historical behavior and is probably motivated by the fact that QTableView does not automatically call submit() when editing leaves a row. The obvious shortcoming of this is that the last row to be edited will not be submitted automatically. It also prevents us from flagging rows other than the pending row as read-only. For OnRowChange, setRecord(), being row-oriented by nature, should submit the change immediately rather than waiting for the next call to setRecord(). This makes setRecord() consistent with insertRecord(). Change-Id: Icb4019d8b7c53a7ee48f8121a7a525e8bc35d523 Reviewed-by: Honglei Zhang <honglei.zhang@nokia.com>
2012-02-28 22:41:31 +00:00
-setRecord() now automatically submits for OnRowChange.
* QSqlQueryModel::indexInQuery() is now virtual. See
QSqlTableModel::indexInQuery() as example of how to implement in a
subclass.
* QSqlTableModel::selectRow(): This is a new method that refreshes a single
row in the model from the database.
* QSqlTableModel edit strategies OnFieldChange/OnRowChange QTBUG-2875
Previously, after changes were submitted in these edit strategies, select()
was called which removed and inserted all rows. This ruined navigation
in QTableView. Now, with these edit strategies, there is no implicit select()
done after committing. This includes deleted rows which remain in
the model as blank rows until the application calls select(). Instead,
selectRow() is called to refresh only the affected row.
* QSqlTableModel::isDirty(): New overloaded method to check whether model
has any changes to submit. QTBUG-3108
QSqlTableModel::setData()/setRecord(): fix incorrect row For OnFieldChange and OnRowChange, before submitting new changes, setData() and setRecord() attempt to submit pending changes and revert them upon failure. However, they fail to consider that reverting pending insertions removes rows from the model. As a result, the new change can be applied to a row higher than intended. One possible solution would be to adjust the targetted index for the removed rows, so that the intended row is affected by the new change. But this still causes the strange editing experience as rows jump up just as they are being edited. It does not seem right in the first place for the model to initiate reverting changes. It should be up to the application to decide what to do when data cannot be committed. In particular, setData() and setRecord() should not have the side effect of reverting already pending changes. The chosen solution is simply to refuse new changes that don't make sense for the edit strategy. For OnFieldChange, flag() will indicate read-only when editing is blocked by a pending change. Since setData() and setRecord() submit data immediately for OnFieldChange, it no longer makes sense to resubmit changes automatically before a new change. For OnRowChange, setData() keeps the behavior of automatically submitting a pending row before starting on a new row. This is historical behavior and is probably motivated by the fact that QTableView does not automatically call submit() when editing leaves a row. The obvious shortcoming of this is that the last row to be edited will not be submitted automatically. It also prevents us from flagging rows other than the pending row as read-only. For OnRowChange, setRecord(), being row-oriented by nature, should submit the change immediately rather than waiting for the next call to setRecord(). This makes setRecord() consistent with insertRecord(). Change-Id: Icb4019d8b7c53a7ee48f8121a7a525e8bc35d523 Reviewed-by: Honglei Zhang <honglei.zhang@nokia.com>
2012-02-28 22:41:31 +00:00
* QSqlTableModel::setData() and setRecord() no longer revert pending changes
that fail upon resubmitting for edit strategies OnFieldChange and OnRowChange.
Instead, pending (failed) changes cause new changes inappropriate to the
edit strategy to be refused. The application should resolve or revert pending
changes. insertRows() and insertRecord() also respect the edit strategy.
QSqlTableModel::setData()/setRecord(): fix incorrect row For OnFieldChange and OnRowChange, before submitting new changes, setData() and setRecord() attempt to submit pending changes and revert them upon failure. However, they fail to consider that reverting pending insertions removes rows from the model. As a result, the new change can be applied to a row higher than intended. One possible solution would be to adjust the targetted index for the removed rows, so that the intended row is affected by the new change. But this still causes the strange editing experience as rows jump up just as they are being edited. It does not seem right in the first place for the model to initiate reverting changes. It should be up to the application to decide what to do when data cannot be committed. In particular, setData() and setRecord() should not have the side effect of reverting already pending changes. The chosen solution is simply to refuse new changes that don't make sense for the edit strategy. For OnFieldChange, flag() will indicate read-only when editing is blocked by a pending change. Since setData() and setRecord() submit data immediately for OnFieldChange, it no longer makes sense to resubmit changes automatically before a new change. For OnRowChange, setData() keeps the behavior of automatically submitting a pending row before starting on a new row. This is historical behavior and is probably motivated by the fact that QTableView does not automatically call submit() when editing leaves a row. The obvious shortcoming of this is that the last row to be edited will not be submitted automatically. It also prevents us from flagging rows other than the pending row as read-only. For OnRowChange, setRecord(), being row-oriented by nature, should submit the change immediately rather than waiting for the next call to setRecord(). This makes setRecord() consistent with insertRecord(). Change-Id: Icb4019d8b7c53a7ee48f8121a7a525e8bc35d523 Reviewed-by: Honglei Zhang <honglei.zhang@nokia.com>
2012-02-28 22:41:31 +00:00
* QSqlTableModel::setData() and setRecord() in OnRowChange no longer have the
side effect of submitting the cached row when invoked on a different row.
****************************************************************************
* Database Drivers *
****************************************************************************
****************************************************************************
* Platform Specific Changes *
****************************************************************************
Qt for Linux/X11
----------------
Qt for Windows
--------------
* Accessibility framework uses IAccessible2
Qt for Mac OS X
---------------
Qt for Embedded Linux
---------------------
Qt for Windows CE
-----------------
****************************************************************************
* Compiler Specific Changes *
****************************************************************************
****************************************************************************
* Tools *
****************************************************************************
- Build System
* Remove qttest_p4.prf file. From now on we should explicitly enable the
things from it which we want. Autotest .pro files should stop using
'load(qttest_p4)' and start using 'CONFIG+=testcase' instead.
- Assistant
- Designer
* [QTBUG-8926] [QTBUG-20440] Properties of type QStringList now have
translation attributes which apply to all items.
They are by default translatable.
- Linguist
- rcc
- moc
* [QTBUG-20785] The moc now has a -b<file> option to #include an additional
file at the beginning of the generated file.
- uic
- uic3
- qmake
* QMAKE_MOC_OPTIONS variable is now available for passing additional parameters
to the moc.
- configure
- qtconfig
****************************************************************************
* Plugins *
****************************************************************************
- The text codecs that were previously plugins are now built into QtCore.
****************************************************************************
* Important Behavior Changes *
****************************************************************************
- QPointer
* QPointer itself is now deprecated, and the implementation of QPointer
has been changed to use QWeakPointer. The old guard mechanism has been
removed. This causes two slight changes in behavior when using
QPointer:
* When using QPointer on a QWidget (or a subclass of QWidget), previously
the QPointer would be cleared by the QWidget destructor. Now, the QPointer
is cleared by the QObject destructor (since this is when QWeakPointers are
cleared). Any QPointers tracking a widget will NOT be cleared before the
QWidget destructor destroys the children for the widget being tracked.
* When constructing a QSharedPointer to take ownership of an object after a
QPointer is already tracking the object. Previously, the shared pointer
construction would not be affected by the QPointer, but now that QPointer
is implemented using QWeakPoiner, constructing the QSharedPointer will
cause an abort().
- QVariant
* Definition of QVariant::UserType changed. Currently it is the same as
QMetaType::User, which means that it points to the first registered custom
type, instead of a nonexistent type.
- QMessageBox
* The static function QMessageBox::question has changed the default argument
for buttons. Before the default was to have an Ok button. That is changed
to having a yes and a no button.