A variadic version of the QSKIP macro has been left in qtestlib for
a transitional period to give upstream projects time to catch up with
the removal of the function from the API. This transitional period is
now finished.
Task-number: QTBUG-21652
Change-Id: Ib1a9158c2efbae05ffd0e69e72879f6387d4b2e8
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
Testlib no longer does anything with the test location, and neither do
any of Qt's tests, so the code is no longer needed.
Change-Id: I1a464116179f3e26efadad6901f00f3c33e0ce84
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
This patch improves performance when constructing a custom type in
a QVariant by ~ 7-20% (instructions count) depending on the type size
and metatype attributes.
Change-Id: Ic2707ff5abd689b66e23c1794f111504bf9b3b01
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
Currently QMetaType API contains almost only static methods. This works
nice until someone needs more information or needs to do more operations
on a type. In this case every function call has to do type dispatch.
This API allows to avoid redundant type dispatching, by caching a
type information in a QMetaType instance. It gives significant
performance boost especially for custom types (up to 9x).
Change-Id: I223d066268402e072e41ca1d0a3e7bc160655d7f
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
This allows us to catch spurious libjpeg warnings via the installation of a
message handler
Change-Id: I82f3257118b5fbdf66550be80e72f75e9a24d3c1
Reviewed-by: Girish Ramakrishnan <girish.1.ramakrishnan@nokia.com>
Reviewed-by: Johannes Zellner <johannes.zellner@nokia.com>
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
-Some of the public classes were missing the \inmodule command.
-Fixed a sentence.
Change-Id: I88ebe12680c744e32253dc01c5ddb4292267caf9
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
The function was already marked as internal.
By moving access to the underlying file engine to private API this
supports the effort to completely drop file engines from public API. In
the future, the goal is to completely drop the file-engine abstraction
as it exists today.
Change-Id: I332fa56e70c87e83c1e08bb9f75e04df7c93fec7
Reviewed-by: Jonas Gastal <jgastal@profusion.mobi>
Reviewed-by: Anselmo L. S. Melo <anselmo.melo@openbossa.org>
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
openStandardStreamsBufferedStreams would fail if standard streams were
redirected to a file (e.g. ./tst_qfile > testlog.txt).
openStandardStreamsFileDescriptors already has a workaround, so apply it
here too.
Change-Id: Iffe9d7864909e489e77c1114e80c4e3bc70a8722
Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
Do like moc: If the string has already been entered into the table,
just return its position, don't make a new copy. This can save
space, for example, if there are several properties of the same type;
the typename only occurs once in the string table but will be
referenced by several property descriptors.
Change-Id: Ic0087697716cab1c6449ea51c0c758a6fd1a1c82
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
Reviewed-by: Kent Hansen <kent.hansen@nokia.com>
14 is the number of fields (ints) in the QMetaObjectPrivate struct
as of revision 6.
Use the calculated number of fields instead, so that the code will
still be correct when more fields are added in future revisions.
Change-Id: I4f2c2bfc125f3fabc8e8caedf5c6ba6c17a34d06
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
buildMetaObject() can operate in two "modes", based on whether a
buffer to write the meta-object into is passed or not. Add asserts
to make sure that the intermediate meta-data indexes are correct
in both "modes", and that the final size in "write mode" matches
the size that was computed in the preceding non-writing pass.
The asserts make it easier to catch obvious problems when changing
buildMetaObject() to generate a new meta-object revision.
Change-Id: Ief7c74e6f6fca836587e831b06072d6aa98c7193
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
Reviewed-by: Kent Hansen <kent.hansen@nokia.com>
Instead of refcounting QWindow visibility, we ask the Application
subclass whether quitting is appropriate.
Task-Id: QTBUG-24120
Change-Id: Idd19cc1a3e5742fddded89c7638aaaa5e47c568d
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
it is *ugly* to have the binary in the repository.
this adds a few seconds to the windows build, as the configure needs to
be rebuilt, obviously. that's almost negligible.
Change-Id: I40ffde23b3c3af2b6bab3e78cd0a9f433214b563
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
We should not assume that the first type id is 0.
Change-Id: I17ba6ba57e97ebd495904bfd11235fe458f214e5
Reviewed-by: Kent Hansen <kent.hansen@nokia.com>
- Pass it from QWindowSystemInterface via QWindow to the widgets.
- Add handler code from 4.8 / qapplication_win.cpp to qwidget.cpp.
Change-Id: Ic759563aa00cb93fe014c1bf41020446c1927dec
Reviewed-by: Morten Johan Sørvig <morten.sorvig@nokia.com>
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
Change-Id: I6858c324548373c57963b5ef137772a1f780ec78
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
Reviewed-by: Clinton Stimpson <clinton@elemtech.com>
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
The patch also speeds up swapSections and hideSection a lot.
It work by eliminating the Span. (That is forcing each 'Span'
to have exactly one element) That saves a lot of loops since
we can often lookup info fast - and/or change it fast.
Since it is often a complexity change, it is difficult to
put %-increase on. (The most used linear function is
recalcSectionStartPos() - and it has a very low constant)
However comparing with the new benchmark (2500 rows)
swapSection, showHideSection and moveSection are about
20-40 factors faster. (Yes, it is a lot faster!)
In the benchmark moveSection is about 300 factors faster.
Beside being a far better model it is also far more simple.
This fix partly solves:
Task-number: QTBUG-19092
Change-Id: I8deeb9315276d15c68e8a27d5dcb8e0c0badf367
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
There are no rights for typical user or even administrator to write
directly under c:\ root in windows without rights elevation, so
completionOnLevelAfterRoot() test case failed. Changed the test case
to use an existing directory.
Task-number: QTBUG-24289
Change-Id: I6a8dfc9d1d6ae798b3b9049c542b45fdbdbd9a8c
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
The XFAILed initStyleOption() case passes on Windows, so do not XFAIL
it there.
Task-number: QTBUG-24297
Change-Id: I9615c408aa7e72b5eb8fe9739903594e45eb5fd7
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
moc can now embed meta information about the plugin
inside the plugin itself. This information can
be queried by Qt without having to load the plugin.
Source compatibility with the old plugin loading
mechanism is still there, but will be removed before
Qt 5.0.
Change-Id: I03e4196ddfed07d0fe94acca40d5de8a6ce7f920
Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The JSON support will get used in moc to support
the creation of plugin metadata that's embedded
into the plugin itself.
Change-Id: I3bc52b16ca0a43bc8bf9141b450045c6183b7823
Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This autotest checks that meta-methods can be properly inspected
(signature, return type, parameter types, etc.).
Change-Id: I13dc75ec5123280e94ec738dade3f54e427fdbaa
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
Removed a couple of header entries since they were causing build
problems on Mac. Additionally it is a bit much to include QtGui/QtGui
in a private header file. Added in an include to QCache and QGradient
which was necessary afterwards.
Change-Id: I442efc799a4c81b50ec30d58f1aedfb08e686fcb
Reviewed-by: Samuel Rødal <samuel.rodal@nokia.com>
As the documentation says, this can be useful to free or recreate
attachments when needed. For example, it might be useful to free stencil
and depth attachments to free up resources when not rendering to the
framebuffer object.
Change-Id: Ib267024fdd380a788c256eb8fb86e0f8832329e0
Reviewed-by: Kim M. Kalland <kim.kalland@nokia.com>
- Run each test in a temporary directory, avoid writing test
files in source/build tree and prevents tests being influenced
by left-overs from previous runs and locks of the application
on the current directory.
- Modify test to be able to use absolute paths to the temporary
directory.
- Skip parts of test removeFileAndUnWatch if a race condition
occurs.
Task-number: QTBUG-24029
Change-Id: I215cc2e0fe6f92d2ffe597b01cdc9c9a39e3c5b4
Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
Do not mark with insignificant_test anymore.
Task-number: QTBUG-22747
Change-Id: I4ef6d5d7e1189b03fd1ab812a0839e3709686e1b
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
For template-based connect(), the meta-object is resolved at
compile-time (the virtual metaObject() function isn't called).
But we can make it work by copying the members of the dynamically
constructed meta-object to the statically defined one.
Change-Id: Ia4d3263a89008e36e187c584db6d25d9042f32b3
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
moc supports it, so qmetaobjectbuilder should too.
Change-Id: I01475794e928b5a1b659f0dab044933948186971
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
Reviewed-by: Kent Hansen <kent.hansen@nokia.com>
Some headers and constants are available in MinGW-w64 that are not
available in the official MinGW. STATE_SYSTEM_HASPOPUP and
STATE_SYSTEM_PROTECTED constants are defined if they are not already
defined by including oleacc.h. _CrtSetReportMode is not used and
crtdbg.h is not included when using official MinGW as crtdbg.h is
missing from official MinGW.
Change-Id: Ie7f3f3726a1663d0fdeb6ee17b86873ae3f61860
Reviewed-by: Jonathan Liu <net147@gmail.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
Should return invalid QModelIndex since inserted row does not map to
query.
Change-Id: Ib1d15cf4198a7063717fb3f3b594b2b1d8a54dfe
Reviewed-by: Yunqiao Yin <charles.yin@nokia.com>
It's not appropriate to use indexInQuery() here. First of all,
the row might be an inserted row, and thus not be in the query.
The intent was probably to get the column position in the query,
but this is certainly not row dependent. Furthermore, if there
are inserted or removed columns, these are managed within
QSqlQueryModel.
Change-Id: I89668655b263747a5b849136404112e911722b3d
Reviewed-by: Yunqiao Yin <charles.yin@nokia.com>
Consider what happens the 1st change succeeds and the 2nd fails. No
select will be done. When submitAll() is called again, the 1st will
still seem to be pending. It will fail or have unexpected effects if
the primary values were changed.
The solution is to avoid resubmitting successful changes. We leave
them in the cache so they stay visible. Submitted changes cannot
be reverted of course.
Change-Id: Ibf400555effa1c3801d02f8713b4b69856ede23a
Reviewed-by: Yunqiao Yin <charles.yin@nokia.com>
They can be generated on demand regardless of edit strategy.
Change-Id: I1e1853e93cc453f1486b65ce577f00141b9c5c47
Reviewed-by: Yunqiao Yin <charles.yin@nokia.com>
There is no reason for the caller to be concerned with this.
primaryValues() now takes advantage of the fact that QSqlQueryModel
uses indexInQuery which was recently made virtual.
Change-Id: I7d856ee05f55c3199fd17c618e559320d0582989
Reviewed-by: Yunqiao Yin <charles.yin@nokia.com>
The purpose of the hack was to fool QSqlQueryModel into signaling the
removal of extra rows via rowsRemoved(). The extra rows are the
inserted rows generated by QSqlTableModel.
While it is important to signal the removal of all the rows before
requerying after committing changes, there is a cleaner way. The
table model should remove its rows before the query model removes its
rows.
Iterating backwards avoids having to decrement row numbers above ones
being removed.
Expected test results have been adjusted for these changes.
Change-Id: I0e8aa81f5e7b8fea5922f5ffd1cfb4a932313a10
Reviewed-by: Yunqiao Yin <charles.yin@nokia.com>
Reading STL iteration code is painful enough if you only have
to do it once.
Thiago suggested remembering the end iterator for performance.
Change-Id: Ic2cdc480f591932ea420e692a4d2796d49f05313
Reviewed-by: Yunqiao Yin <charles.yin@nokia.com>
Qt 5 seems like an excellent opportunity to simplify logic and separate
concerns by making indexInQuery() virtual. Note that this wasn't my
idea, but was mentioned in a helpful comment.
Change-Id: Ie29ead110def45297c32de3ce6d07a8eefb08d8c
Reviewed-by: Yunqiao Yin <charles.yin@nokia.com>
QLocalServer could only listen to sockets it created.
Thi is not always possible as sockets may be passed
by socketpair() or have to be created locally by
other means. This adds a similar feature to QLocalSocket
where a native descriptor maybe used.
Change-Id: I43b0af179b3b868dd164d4e1fd312ff4546cf9ff
Reviewed-by: Michalina Ziemba <michalina.ziemba@nokia.com>
Reviewed-by: Tapani Mikola <tapani.mikola@nokia.com>
Reviewed-by: Harald Fernengel <harald.fernengel@nokia.com>
The fileLineEndingTest case doesn't test network in any way and it is
conceptually wrong, too, as any tests where line endings are an issue
should be handled with .gitattributes rather than forcing user to
check out the repo with unix line endings.
Task-number: QTBUG-24271
Change-Id: I73986993edc227cb68b8f61d51cc1cf458d20989
Reviewed-by: Shane Kearns <ext-shane.2.kearns@nokia.com>
Sometimes GPOS tables are used for kerning and can grow quite
large if the font is unoptimized. In that case, allocating the
maximum size for each ValueRecord will have a great impact on
memory consumption. For GPOS tables which do not contain device
tables, we only allocate the memory actually need to store the
data instead.
While it would be possible to optimize memory for GPOS tables that
contain device tables as well, this is not a part of this patch.
Change-Id: Id665b2821675ef955c497c782f09f99af765b8a3
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
