diff options
author | Jan Arve Sæther <jan-arve.saether@qt.io> | 2020-02-04 11:32:00 +0100 |
---|---|---|
committer | Jan Arve Sæther <jan-arve.saether@qt.io> | 2020-02-06 09:31:47 +0100 |
commit | 8c6c4df3e83776d821f357582a717dbfbeb1c3ff (patch) | |
tree | aaeb78842920fcffd3dcb6ec2d62643c0557ce64 /tests/auto/widgets/widgets/qtabbar | |
parent | 6d6ca70538851e8334eb01550a93b60ceba75d34 (diff) |
Maintain at least 500ms timestamp distance between each test function
If we had one test function that just did
tst_Mouse::f1()
{
QTest::mouseMove(w, QPoint(0,0));
}
and another test function that did
tst_Mouse::f2()
{
QTest::mouseMove(w, QPoint(500,500));
}
their corresponding event timestamps were only 1 apart from each other.
This meant that any code that tried to estimate the velocity of a mouse
cursor would get a really high velocity estimate inside f2(). This would
come as a surprise to most people. So to avoid this, we add a 500 ms
timestamp delay between each test function call.
In theory this could also prevent generating a mouseDoubleClickEvent
when a pair of test functions containing a press-release sequence was
run, but there is a separate pre-existing mechanism to handle that case.
Change-Id: Icd4fc35853c09f080466d22411208c7b5c4174b5
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Diffstat (limited to 'tests/auto/widgets/widgets/qtabbar')
0 files changed, 0 insertions, 0 deletions