Disable the crash signal handler. This makes it possible
to run auto-tests in the browser.
Long-running tests may cause the browser to interrupt
or display the “a web page is slowing down your computer”
message, or not produce any console output while the
test is running.
Change-Id: Ifd53b744bd3652abfb466b78992ce2371eca2536
Task-number: QTBUG-68504
Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
This is a temporary measure to work around an implementation bug on
Integrity: For all other platforms, QtPrivate::condition_variable is
just std::condition_variable. On Integrity, it's a class that wraps
QWaitCondition to provide the interface of std::condition_variable.
This allows the use of std::condition_variable across Qt without
running into the Integrity issue. Once we can depend on an more modern
Integrity toolchain, removing QtPrivate::condition_variable is a
simple mechanical change:
s/QtPrivate::condition_variable/std::condition_variable/g;
s/QtPrivate::mutex/std::mutex/g;
Task-number: QTBUG-78450
Change-Id: I293a99d1cdc48691817b926aa51ecd84556e5e90
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
... by porting from (QWaitCondition,QMutex) to std::{condition_variable,
mutex} and making the code more accessible by introducing explicit
states.
This patch was originally starting out to just replace QWaitCondition
with std::condition_variable, which is faster and more compact, and
the patch still does that, too. The focus, however, has shifted
towards improving the accessibility of the code, in particular its
states and the transitions between them.
Due to the way QWaitCondition prevents spurious wakeups, this code
could use its QWaitCondition as a semaphore, e.g. where it was calling
wait() in the ctor to be released only when the thread it started has
entered run(). That makes the code unnecessarily hard to follow.
Fix by introducing an enum Expectation which tells what the watch-dog
is currently waiting for, and unmistakably determines the state the
watch-dog is in: ThreadStart, TestFunctionStart, ... The timeout value
is now selected based on the Expectation value, so the timeout is no
longer serving as the implicit state, either.
Also port the defaultTimeout() function to return a
std::chrono::milliseconds directly. Elsewhere in Qt, we guard against
lack of <chrono>, so this unguarded use will also serve as a guide to
see whether all supported platforms now, eight Qt releases after we
formally require C++11 for Qt usage, provide it. Without <chrono>,
there's no timed waiting in the standard library.
Change-Id: If97b601c4090a2a2926fa58c903cfe3ec2656324
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Add overloads for qCompare() for QStringView making use of the fast
formatting helper introduced by 94aa350621
for int.
Speeds up the bug report example by a factor of 3..4.
Task-number: QTBUG-38890
Change-Id: Icc706618b2f1d23b37d354a04d4e1d1cc4b5aee3
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
- Fix else after return/break
- Fix use of int as boolean literal
- Use range-based for in some cases
- Avoid copies by using const-ref
- Remove unnecessary null-check before delete
Task-number: QTBUG-69413
Change-Id: I69f46e6deaa55ef70a8b3a61e6539c79a64aaa23
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Apply Fixits by Qt Creator with some amendments.
Task-number: QTBUG-69413
Change-Id: I620e40a277dc2b20b0ec26fc32577e66b0456bb3
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
Factor out a helper template formatting the QCOMPARE failure message
delaying the formatting of the parameters with toString() and use that
for float, double, int and unsigned. This removes the need to always
format and allocate strings for the operands even in the success case,
speeding up the QColor test from 3.3s to 700ms (Windows/release).
Task-number: QTBUG-38890
Change-Id: I999484765bdaed921d3fc35f35a9fbbcd82a9704
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
If the resource path 'base' was relative it would enter condition 3) in
qFindTestData() and it would actually find the nonsensical "://data" as a
viable candidate. We don't want to enter that case, but rather enter the
subsequent ('Try resources') case
Change-Id: I1928ba02c941e23fee4fec9052a1981e46fa59b7
Task-number: QTBUG-73512
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Apply Fixits by Qt Creator with some amendments.
Task-number: QTBUG-69413
Change-Id: I366cca6e5755719e8241e76774af6be2b5312627
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Semi-automated, just needed ~20 manual fixes:
$ find \( -iname \*.cpp -or -iname \*.h \) -exec perl -pe 's/(\.|->)load\(\)/$1loadRelaxed\(\)/g' -i \{\} +
$ find \( -iname \*.cpp -or -iname \*.h \) -exec perl -pe 's/(\.|->)store\(/$1storeRelaxed\(/g' -i \{\} +
It can be easily improved (e.g. for store check that there are no commas
after the opening parens). The most common offender is QLibrary::load,
and some code using std::atomic directly.
Change-Id: I07c38a3c8ed32c924ef4999e85c7e45cf48f0f6c
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
- braces on same line as else where applicable
- const where possible
- Same extended logging format for all cases
Change-Id: If0c91f270ce2a9be1d295f42a915900f3b1838cf
Reviewed-by: Christian Stenger <christian.stenger@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
We were potentially adding the Apple test logger multiple times, and we
didn't consider whether the existing loggers were logging to file or not
when circumventing them.
We now don't use the Apple logger if it would touch stderr and some other
logger is using stdout. In the case of no explicit logger being specified
on the command line, we allow the Apple logger to take priority over the
default plain test logger.
Change-Id: I31bbec4f4b3ab84ba9a2be35e8e5db08fee071a7
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Tidied up the existing float tests in the process.
(In particular, s/SUCCESS/PASS/ since that matches real test output.)
These verify that QCOMPARE() handles floats and doubles as intended.
Extended the existing qFuzzyCompare tests to probe the boundaries of
the ranges of values of both types, in the process.
Revised the toString<double> that qCompare() uses to give enough
precision to actually show some of the differences being tested there
(12 digits, to match what qFuzzyCompare tests, so as to show different
values rather than, e.g. 1e12 for both expected and actual) and to
give consistent results for infinities and NaN (MinGW had eccentric
versions for these, leading to different output from tests, which thus
failed); did the latter also for toString<float> and fixed stray zeros
in MinGW's exponents (which made a kludge in tst_selftest.cpp
redundant, so I removed that, too).
That's further complicated handling of floating-point types, so let's
just keep an eye on how expensive that's getting by adding a benchmark
test for QTest::toString(). Unfortunately, default settings only get
runs that take modest numbers of milliseconds (some as low as 40)
while increasing this with -minumumvalue 100 or more gets the process
killed - and I'm unable to find out who's doing the killing (it's not
QProcess::kill, ::kill or the QtTest WatchDog, as far as I can tell).
So results are rather noisy; the integral tests exhibit speed-ups by
factors up to 5, and slow-downs by factors up to 100, between runs
with and without this change, which does not affec the integral tests.
The relatively modest slow-downs and speed-ups in the floating point
tests thus seem likely to be happenstance rather than signal.
Change-Id: I4a6bbbab6a43bf14a4089e96238a7c8da2c3127e
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Based on Asmo Saarela's advice (QTPM-686), adapted on advice from
FrogLogic support and converted to a feature so that the selftest and
testlib qmake config can be co-ordinated.
Task-number: QTPM-1385
Change-Id: Icd706f086009e1e08b3f8c5cd553f792402e28c0
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
The qCompare() implementation for double was handling infinities and
NaN the way tests need, but the one for float didn't; it has just the
same need, so apply the same fix. Extends 79493a3ee1.
Change-Id: I8425026acb61d535e449f579b77fdcd609157f7c
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
It's not actually exercised by anything in our source tree, but is
potentially useful and has been part of the documented public API for
some time. So mention that the caller is responsible for delete[]ing
its return and add a test that exercises it.
Task-number: QTPM-1385
Change-Id: Ifc5284b9eb1b678cf3c9708c681311e874838fc6
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@qt.io>
Use array dereferencing notation ptr[expr] rather than the more
verbose and less readable *(ptr + expr).
Change-Id: Ie98986e7f622aa1a94f3f14bc362ed65767f9d92
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@qt.io>
QTest::qtest_qParseArgs() takes a needlessly mutable parameter.
This needs transient legacy version until qtdeclarative is converted
to use the const-qualified variant.
Change-Id: I19fad25423f6f1ea7d5fac537815f4f633ac4405
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@qt.io>
This reverts 4fe68ffbe5,
1dfc16f6da and
9545bee98a2431d4ef4c3449631a5fcb8d9bd57a; this feature has not been in
use for some time, so is just dead code.
[ChangeLog][QtTestLib] Removed support for GPU_BLACKLIST files.
Reimplementations or equivalents of QTEST_MAIN() should remove their
uses of QTEST_ADD_GPU_BLACKLIST_SUPPORT and its _DEFS; they are still
vacuously defined, but serve no remaining purpose and shall be
undefined in due course.
Change-Id: I94ffd5c37ce4e1f7cf25d3c7ae8d40696b74c911
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Add it to configure.json and replace all occurrences of QT_NO_THREAD
with QT_CONFIG(thread). Add conditions for other features that depend
on thread support. Remove conditions where we can use the QMutex and
QThreadStorage stubs.
Change-Id: I284e5d794fda9a4c6f4a1ab29e55aa686272a0eb
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Makes it easier to run a test repeatedly until it crashes, and then
attach a debugger.
Change-Id: I8ad5d706becd77a2743a51927c837f3d1da08624
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Having the helpers in each respective module lets us implement the
helpers using private APIs without forcing the test to add private
dependencies.
It also makes it easier to test Qt using a third party testing
framework (for running the test suite), while still using the helpers
for ensuring tests behave expectedly.
Change-Id: I2a6ce24526ed345f3513548f11da05c7804c203f
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Fix warnings about invalid function type casts (return types
conflicting with the FARPROC returned by GetProcAddress()) like:
corelib\global\qoperatingsystemversion_win.cpp💯48: error: cast between incompatible function types from 'FARPROC' {aka 'long long int (*)()'} to 'RtlGetVersionFunction' {aka 'long int (*)(_OSVERSIONINFOW*)'} [-Werror=cast-function-type]
io\qlockfile_win.cpp:158:85: error: cast between incompatible function types from 'FARPROC' {aka 'long long int (*)()'} to 'GetModuleFileNameExFunc' {aka 'long unsigned int (*)(void*, HINSTANCE__*, wchar_t*, long unsigned int)'} [-Werror=cast-function-type]
by introducing nested casts.
Task-number: QTBUG-68742
Task-number: QTQAINFRA-2095
Change-Id: I3a5d2ea901bf5dc35963c589d61cf3dc7393377a
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
When cross-compiling for Android on Windows the 'relative to test
source' option could end up possibly matching with a matching folder in
root ('/') because file is a Windows path and the canonicalFilePath of
that path is "".
Fixes tst_qxmlstream (and possibly others) on Android when Qt is
compiled on Windows.
Task-number: QTBUG-68596
Change-Id: I378374e41eea80f43680b3941adaa91fa604934a
Reviewed-by: Jesus Fernandez <Jesus.Fernandez@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Currently when doing comparison with std::tuple the fallback toString
method is called which returns a Q_NULLPTR thus not allowing proper
diagnostic of the values that triggered an error. This patch
adds support for std::tuple to improve the tests output readability.
[ChangeLog][QtTest][QCOMPARE] Now outputs contents of std::tuple on
failure.
Change-Id: I046a55e2ce44c3f7728d51e4745120d38aa5e007
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This will make two floating points containing NaN compare as equal,
instead of the regular nan != nan IEEE behavior (which isn't very useful
in a unit-test framework).
Note that this does not apply to indirect comparisons, for example via
QVariant.
Change-Id: I39332e0a867442d58082fffd150851acfdd18c23
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
This helps to distinguish timeouts from real crashes in COIN.
This is already done for Windows.
Change-Id: I4daeafa36f50482d20cea4bd1106647081ff7abe
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>
Use QT_CONFIG(regularexpression), disentangle it from QT_BOOTSTRAPPED,
switch it off in the bootstrap build, remove the #ifdefs from
qregularexpression.{h|cpp}, and add QT_REQUIRE_CONFIG(regularexpression)
to the header.
qregularexpression.{h|cpp} are already correctly excluded in tools.pri
if !qtConfig(regularexpression).
Change-Id: I21de154a6a118b76f99003d3acb72ac1e220d302
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
The Test Anything Protocol (TAP), was originally Perl's simple text-based
interface between testing modules and test harnesses, but has since been
adopted by a large number of producers and consumers in many different
languages, which allows colorizing and summarizing test results.
The format is very simple:
TAP version 13
ok 1 - test description
not ok 2 - test description
---
message: 'Failure message'
severity: fail
expected: 123
actual: 456
...
ok 3 - test description # SKIP
1..3
The specification [1] is very brief, so the implementation has been
based on how typical consumers behave, especially when it comes to
the undefined diagnostics block.
[1] http://testanything.org/tap-version-13-specification.html
Change-Id: I616e802ea380165c678510e940ddc6607d39c92d
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
After a37785ec76 went in, it become
apparent that multi mouse button state handling in qtestlib is
non-existent, for details see QTBUG-64030 and QTBUG-63786. What
happened behind the scenes often was not what one would expect based
on the provided QTest::mouse* input sequence - events went missing,
incorrect events were generated, each subsequent test function
started with a state set from the function that run earlier. It is
easy to see how a minor change in one test could easily affect outcome
of other tests.
With a37785ec76, Qt platform plugins
are now responsible for sending explicit mouse button type and state
information; qtestlib should take full responsibility now as well.
But using the new API from a37785ec7 alone in qtestlib is not sufficient.
We need to reset mouse state between each new test function run (we do
this at function scope as that fits with the current qtestlib API user
expectations). This patch implements the necessary reseting logic.
Updated tst_qwindow.cpp::generatedMouseMove() to use QTest::mouse* APIs.
That test requires pressing multiple buttons, it was not possible with
QTest::mouse* APIs before this patch.
Added an auto test for multiple mouse button pressing/release in
tests/auto/testlib/selftests/mouse/. And few other tests which are
currently QSKIP-ed, but should be considered when re-designing qtestlib
APIs.
Task-number: QTBUG-64030
Change-Id: I39fdcbc73a467a7463ce2aed622bf22484095635
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Qt Forward Merge Bot
Morten Johan Sørvig
Liang Qi
Marc Mutz
Allan Sandfeld Jensen
Friedemann Kleint
Jan Arve Sæther
Giuseppe D'Angelo
Oliver Wolff
Tor Arne Vestbø
Robert Loehning
Edward Welbourne
Nico Vertriest
Tasuku Suzuki
Ulf Hermann
Mårten Nordheim
Samuel Gaist
Thiago Macieira
Gatis Paeglis