Restructured, separated the canonicalising of output lines out (into
an object that prepares the necessary regexes and replacements)
suppress more, changed the path-stripping to strip qtbase's parent
rather than os.getcwd() and took account of shadow builds (so both
source tree and build tree provide prefixes we want to strip from
paths). Also cope with $PWD potentially having symlinks in it, where
os.getcwd() is canonical.
It's possible some output might name files elsewhere in the source
tree; these won't be filtered by the prior cwd prefix removal; and, in
any case, the problem with cwd is only that the ancestry of qtbase is
apt to vary; paths relative to there should be consistent between test
runs. This change shall lead to a one-off rewrite of all expected_*
files; but it should now catch all paths. By stripping both build
root and source root (when different) it also avoids differences for
those doing out-of-source ("shadow") builds.
In our XML formats, any hyphens in root paths (e.g. I had Qt-5.6 in my
build root's path) got represented by a character entity, confounding
the replacement; so also do replacement that catches this. We may
discover other character entity subsitutions needed along with this.
Now filtering line numbers and timing information, including benchmark
results; these numbers all get replaced with 0 to avoid noisy diffs.
Also purging dangling hspace, to placate sanity-bot.
The module can now be imported - the code it runs is packaged as a
main() function that a __name__ == '__main__' stanza runs - and all
data is localised to where it's needed, rather than held in globals.
Tidied up and organized the existing regexes. There are doc-strings;
there is a short usage comment. Data is localised rather than global
and modern pythonic idioms get used where apt.
Regexes are compiled once instead of repeatedly. An object looks
after the list of patterns to apply and its construction handles all
anticipated problems. Failures are mediated by an exception.
The output file now gets written once, instead of twice (once before
editing, then over-write to edit), and Popen uses text mode, so that
write can do the same. Its command is delivered as an array, avoiding
the need to invoke a shell.
Instead of relying on qmake being in our path (which might give us a
bogus QT_VERSION if the one in path doesn't match our build tree), use
the relative path to qmake - we rely on being run in a specific
directory in the build tree, after all. Escape dots in the version
properly, so that 51730 doesn't get mistaken for 5.7.0 (for example),
and moved this check later in the sequence (matching a smaller target
makes it more likely to falsely match).
Overtly check we are in the right directory and tell the user what we
actually need, if run from the wrong place. Simplify handling of the
unsupported use-case for MS-Windows (but note what would be needed for
it).
Change-Id: Ibdff8f8cae173f6c31492648148cc345ae29022b
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
[ChangeLog][QtGui][QTextDocument] Fixed a bug that would return a wrong
position when searching backward from the end of the document.
Task-number: QTBUG-48182
Change-Id: I6e88f808a50cb840f61e7bc579e2a28c5300089d
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Prevent QAbstractSpinBoxPrivate::interpret() from bailing out
in focus changes after text has been entered.
Task-number: QTBUG-55249
Change-Id: I250b3c50f7db5de2e9356038df20f18ee059df11
Reviewed-by: Maurice Kalinowski <maurice.kalinowski@qt.io>
This makes it possible for the application to control which language
is used by QMimeType::comment()
[ChangeLog][QtCore][QMimeType] QMimeType::comment() now uses the default locale
rather than system locale, so that applications can control which language
is being used.
Task-number: QTBUG-50776
Change-Id: I82623b7c488035a4164fadaf37ebcc79a9fd6173
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Some qpa backends do not provide a QPlatformNativeInterface. Hence,
check whether return value is valid.
Change-Id: Iab46bc59a151aa244fcfebf58edb37496369db89
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
This fixes a problem that occurs when a task, that is run synchronously,
throws an exception. If that happened, then the exception would not be
re-thrown, because of an early return.
Task-number: QTBUG-54831
Change-Id: Ic70c5b810ec6adce6e62bfd6832ba9f170b13a7f
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@theqtcompany.com>
Widgets module is no longer required. Removing it allows this test to
be run with a qtbase configured with -no-widgets, which saves
compilation time.
Change-Id: Id99d3f25cd7b227aa81e1cf1ac7b6fd5227ac4c4
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
As of version 10.12 (Sierra), the name of Apple's desktop operating
system will be macOS. Change the occurrences where the Mac platform
is discussed to use a macro \macos, which expands to 'macOS'. This
helps with adapting to future renaming.
Update the instructions on mac-specific Q_OS_* macro usage.
Add a \target for the old 'Qt for OS X' topic to keep links working
for other documentation modules that try to link with the old name.
Change-Id: Id33fb0cd985df702a4ae4efb4c5fd428e77d9b85
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>
Simply allocate objects on the stack instead of the heap.
Change-Id: Ic047d78e49668878821cce1c8ab599a8551b6476
Reviewed-by: David Faure <david.faure@kdab.com>
In QStringListModel, the display and the edit roles are synonyms,
so when one is changed, the other changes with it. However, in
setData() we only emitted a vector with just the role that was
passed in by the user.
Fix by always passing both roles, regardless of which one was used
to set the data.
Change-Id: I498e7cb33796fae266901817b01ad85d861d4bb4
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The deleteLater() call wasn't reliably reached when tests fail,
so use a QScopedPointer with QScopedPointerDeleteLater deleter.
Change-Id: Ica73bc73c2a0ac1e9b77e4804f2aedcad9b662a0
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@theqtcompany.com>
Reviewed-by: Richard J. Moore <rich@kde.org>
The test no longer writes to SRCDIR, so don't try to remove
generated files from there, either.
Amends bb5570082e.
Change-Id: I1d5df88b1865f3dbd914ec71147de61e173f2f4e
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@theqtcompany.com>
QStyle was created but never removed
Change-Id: I55011377afd475af28e4ce2cf657e435dd37c96a
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
We can't (at present) actually exercise the failure in
QWindowsLocalCodec::convertFromUnicode() that prompted us to consider
the possible failure here, but we should at least test for it.
Change-Id: I5066c88d7b4caeb48aebc6b79c355fa49e1c581c
Reviewed-by: Frederic Marchal <frederic.marchal@wowtechnology.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Sometimes, this test fails in CI due to notifications arriving
asynchronously from the OS. This happens inside closeWriteChannel()
call, where we are flushing the write buffer and I/O completion on
the read pipe could occur there as well. So, take this into account
before waiting for the new incoming data. Also, improve the checks
on successful reading and writing.
Change-Id: Iabe875fc346eb4420c72d03208d22ea861a570c6
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Before: HexRgb: 0.00230 ms per iteration, HexArgb: 0.00290 ms per iteration
After: HexRgb: 0.00051 ms per iteration, HexArgb: 0.00061 ms per iteration
This showed up as a relevant optimization when profiling KIconLoader
which uses QColor::name() as part of the key -- thanks to Mark Gaiser for
the investigation and first suggestion of a solution. I have also seen
customer code writing a replacement for QColor::name() because it was
too slow to be used as a hash key.
Change-Id: I009ccdd712ea0d869d466e2c9894e0cea58f0e68
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Instead of 'true', it should be '-1'.
Change-Id: I5e8f99153da68d34b37477ef4cedbc447fba347f
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Use QStringRef::isNull instead of QStringRef::string()
for validation. Non-NULL str.string() may yet leave us
with a useless str.unicode(), which is the actual problem here;
whereas !str.isNull() does really confirm that str.unicode()
is sensible.
Such test prevents situation like:
const QString a;
QString b;
b.append(a); // b.isNull() == true
b.append(QStringRef(&a)); // b.isNull() == false
Auto test updated: create QStringRef from QString directly, without
any condition.
Change-Id: I082cd58ef656d8a53e3c1223aca01feea82fffb9
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
According to QLocalSocket's documentation, connectToServer() must
initiate a connection attempt after opening the device. Otherwise, if
a connection succeeds immediately, connected() signal will be emitted
on closed device. So, this patch ensures that TCP-based implementation
behaves correctly.
Change-Id: I4cc9474815e091a1491a429a6dc17f9cf0154f58
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
As it were, QStringLists were not handled explicitly when comparing
QVariants. If both QStringLists contained only a single entry, they
were treated as QStrings - if both QStringLists were empty, there were
equal (correctly so) - but if one of the QStringLists had more than
one entry, the compare function fell through to returning always 1.
As discussed here: https://stackoverflow.com/a/38492467/3444217
Added rich comparison tests for all non-numerical, non-recursive
QVariants that support them (except QModelIndex and
QPersistentModelIndex)
Task-number: QTBUG-54893
Change-Id: Icc5480d9ba056ee5efe83da566c5829caa1509d7
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@theqtcompany.com>
As reported by UBSan:
tst_qtreeview.cpp:2187:36: runtime error: downcast of address 0x7ffc15749f20 which does not point to an object of type 'PublicView'
0x7ffc15749f20: note: object is of type 'QTreeView'
Fix by making the test a friend of QTreeView (and, for
Clang, of QAbstractItemView) instead.
Change-Id: I5b748696ab441a91058f4d45a18bd5ed75a6e560
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
As reported by UBSan:
tst_qabstractitemview.cpp:336:23: runtime error: member call on address 0x7ffe6fe96e10 which does not point to an object of type 'TestView'
0x7ffe6fe96e10: note: object is of type 'QListView'
tst_qabstractitemview.cpp:337:5: runtime error: member call on address 0x7ffe6fe96e10 which does not point to an object of type 'TestView'
0x7ffe6fe96e10: note: object is of type 'QListView'
tst_qabstractitemview.cpp:338:23: runtime error: member call on address 0x7ffe6fe96e10 which does not point to an object of type 'TestView'
0x7ffe6fe96e10: note: object is of type 'QListView'
[etc ...]
Fix by making the test a friend of QAbstractItemView instead.
Change-Id: I1a08977042296eb34e9dbdb5c0595662dbd2e5ef
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
We can't forward a VerticalSortHint or HorizontalSortHint hint, because we might
be filtering extra items.
The documentation of QAbstractItemModel::LayoutChangeHint states:
Note that VerticalSortHint and HorizontalSortHint carry the meaning that
items are being moved within the same parent, not moved to a different
parent in the model, and not filtered out or in.
And some of the views rely on this assumption (QQmlDelegateModel for example)
What happens in the test is the following:
- 'model' emit the dataChanged signal when its data is changed.
- 'proxi1' QSortFilterProxyModelPrivate::_q_sourceDataChanged does not forward
the dataChanged signal imediatly, it will instead first re-sort the model and
call layoutAboutToBeChanged / layouChanged with the VerticalSortHint
- 'proxy2' would forward the layoutAboutToBeChanged with the hint, but in
QSortFilterProxyModelPrivate::_q_sourceLayoutChanged, it will redo the mapping
which will cause the changed data to be filtered.
So proxy2 can't forward the VerticalSortHint as it removed rows in the process.
Change-Id: I20b6983e9d18bf7509fe6144c74f37d24e4a18c2
Reviewed-by: Tobias Koenig <tobias.koenig@kdab.com>
Reviewed-by: David Faure <david.faure@kdab.com>
As reported by UBSan:
tst_qtreeview.cpp:663:5: runtime error: load of value 4294967295, which is not a valid value for type 'DragDropMode'
Instead of abusing -1 to indicate to not set the dragDropMode,
use a separate boolean field in tha data.
Change-Id: I13e7539c858f3b2462d57d660062ef3cb7aec61b
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
They look relative because the path doesn't start with a '/' but they
have a scheme so they shouldn't be combined as if it was one absolute
and one relative URL.
[ChangeLog][QtCore][QUrl] QUrl::resolved() no longer treats a URL with
a scheme as a relative URL if it matches this URL's scheme. This special
casing was incompatible with RFC 3986 and broke resolving data: URLs,
for instance.
Change-Id: I3758d3a2141cea7c6d13514243eb8dee5d510dd0
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
There was a race condition between QObject::disconnect() and
QMetaObject::activate() which can occur if there are multiple
BlockingQueued connections to one signal from different threads and
they connect/disconnect their connections often.
What can happen in this case is:
T1 is in activate() method and T2 is in disconnect() method
T1 T2
locks sender mutex
selects next connection
unlocks sender mutex
locks sender mutex
sets isSlotObject to false
creates QMetaCallEvent derefs connection
posts event
Two things can happen here:
1. Connection can still be valid, but it will have isSlotObject==false
and callFunction will be used instead of slotObj
2. Connection can already be invalid
To fix it mutex unlock should be moved after QMetaCallEvent creation.
Also there is another case, when we don't disconnect but delete the
receiver object. In this case it can already be invalid during
postEvent, so we need to move mutex unlock after postEvent.
Change-Id: I8103798324140ee11de5b4e10906562ba878ff8b
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Extract the common part from QItemDelegate and QStyledItemDelegate
which uses QLocale to convert a value for Qt::DisplayRole to a string.
Use this code to get the text for tooltips and "What's this?".
[ChangeLog][QtWidgets][QAbstractItemDelegate] Show localized detailed
tooltips and "What's this?" texts.
Task-number: QTBUG-16469
Change-Id: I8618763d45b8cfddafc2f263d658ba256be60a15
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
It crashed when d was equal to Data::unsharableEmpty().
Task-number: QTBUG-51758
Change-Id: If9f2a7d11892507135f4dc0aeef909f59b7478fc
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@theqtcompany.com>
This is a partial revert of a4e2f2e687.
That fix tried to avoid the risk of a crash in pixel() by ensuring
Mono QImages created with external data also got a default color
table. However, that broke usable behavior in existing code that was
painting in Mono QImages using color0/color1.
This commit reverts to the old behavior, and instead expands on the
checking in pixel() so that lacking color table is handled gracefully
for all indexed formats.
Task-number: QTBUG-54827
Change-Id: I9164198bed9d20c4b12cdba40a31c141bef3128d
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
For platforms not providing mkdtemp(), QTemporaryDir relied on an implementation
of q_mkdtemp() operating on char *, converting back and forth using
QFile::encodeName()/decodeName() when passing the name to QFileSystemEngine.
This caused failures on Windows (which uses "System"/Latin1 encoding)
for names containing characters outside the Latin1 space.
Reimplement q_mkdtemp() to operate on QString, which avoids the conversions
altogether and also enables the use of larger character spaces for the
pattern.
Add tests.
Task-number: QTBUG-54810
Change-Id: Ie4323ad73b5beb8a1b8ab81425f73d03c626d58a
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
We should not call prev() if we had already reched the end.
Task-number: QTBUG-54815
Change-Id: I56bc86880a0dbfdce57fc4a08e5950f2ff3a5958
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@theqtcompany.com>
The XML parser uses fastScanLiteralContent() to read a block of
text. The routine was not checking the range of valid characters as
defined in the XML standard:
https://www.w3.org/TR/2008/REC-xml-20081126/#NT-Char
A check has been added to stop reading the bad character.
Note that the characters are legal in XML 1.1, but QXmlStreamReader
is a well-formed XML 1.0 parser
[ChangeLog][QtCore][QXmlStreamReader]
Fixed a bug in the XML parser that prevented to load XML that
contained invalid characters for XML 1.0.
Change-Id: I10aaf84fbf95ccdaf9f6d683ea7c31925efff36d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Do not increment 'data' past the buffer in case of invalid token.
Remove the left over qDebug so we can make a test.
Task-number: QTBUG-54609
Change-Id: I8f0dd3381fbdea3f07d3c05c9a44a16d92538117
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@theqtcompany.com>
when the file name is empty, the path will be returned verbatim. this
must be considered when constructing the return value.
Task-number: QTBUG-54550
Change-Id: Ie108ed52275e66a154ef63bd6f7193f55b3e0454
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
that can make sense if a function which determines the availability of
a dependency fails to do so for unexpected reasons.
Change-Id: If6cd113df25aee66830c120a2fab067c822a4543
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
if() would simply "downgrade" a fatal error to a false condition, which
is certainly not expected.
Change-Id: Ie9c54f2bddf588856498bf795007b341b7c9363a
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
When setting a QFileSystemModel as model, the completion role
is set to QFileSystemModel::FileNameRole. This needs to be reset
to the default Qt::EditRole when setting another model.
Task-number: QTBUG-54642
Change-Id: Ie78d5d417e008ad05a2f995bdbc218b3ad1bc49c
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
When replacing each copy of one text with a copy of another, we do so
in batches of 1024; if we get more than one batch, we need to keep a
copy of the sought text and replacement if they're part of the string
we're modifying, for use in later batches.
Also do the replacements in full batches of 1024, not 1023 (which left
the last entry in an array unused); marked some related tests as
(un)likely; and move some repeated code out into a pair of little
local functions to save duplcation.
Those new functions can also serve replace_helper(); and it can shed a
const_cast and some conditioning of free() by using them the same way
replace() now does. (There was also one place it still used the raw
after, rather than the replacement copy; which could have produced
errors if memcpy were to exercise its right to assume no overlap in
arrays. This error is what prompted me to notice all of the above.)
Added tests. The last error proved untestable as my memcpy is in fact
as fussy as memmove. The first two tests added were attempts to get a
failure out of it. The third did get a failure, but also tripped over
the problem in replace() itself. Added to an existing test function
and renamed it to generally cover extra tests for replace.
Change-Id: I9ba6928c84ece266dbbe52b91e333ea54ab6d95e
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
In Qt, null QStrings compare equal to empty ones, so add an explicit
check that the corresponding hash values are identical, too.
Ditto for QByteArray.
Change-Id: I190fc95a765305928d9b6b0e4955433865b6b247
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QLineEdit::setText() emits QLineEdit::textChanged(), which is
connected to QComboBox::editTextChanged(). When a user slot
connected to editTextChanged() sets QComboBox::editable to
false, the line edit will be deleted in setEditable() and
when control returns to QComboBoxPrivate::setCurrentIndex(),
the formerly non-null 'lineEdit' has changed to nullptr,
leading to a nullptr dereference when attempting to set the
completionPrefix on lineEdit->completer().
Fix by re-checking 'lineEdit' after returning from the
QLineEdit::setText() call.
Add a test.
Task-number: QTBUG-54191
Change-Id: I94154796cfde73fb490f8471c48b9d6f62825a92
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
These expressions only work because they contain no non-parenthesized
commas and an int literal is last.
Fix by wrapping only the integer literal in Q_(U)INT64_C.
Change-Id: I6b8e508b6c7c022f4b3342f65c26aab89ce17702
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Otherwise, float numbers are formatted using a decimal comma
in German, causing huge diffs and failures.
Change-Id: Icd85a293d0564cac6be244eb0793611920d0c89c
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
