... to make it pass with LLDB. The exact display of the private
is not *that* interesting anyways.
Change-Id: I2d0a8bdd5daff4b03c77b00c887837c7568ea508
Reviewed-by: Christian Stenger <christian.stenger@theqtcompany.com>
LLDB 3.7 on Linux doesn't get the namespace right in QMapNode:
sc t = lldb.target.FindFirstType('Myns::QMapNode<int, CustomStruct>')
sc t.GetName() -> QMapNode<int, CustomStruct> (no Myns::)
Change-Id: Ied306f70035e0669bc822700a92371b7de35a603
Reviewed-by: Christian Stenger <christian.stenger@theqtcompany.com>
Otherwise, if you increase the number of rows again later you get
surprising results.
Change-Id: I3ec809820716368849415f9dcdb1ec9208844353
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
Always add a space before a *. This fixes some test cases with LLDB
that failed to recognize the default argument in "std::vector<int *,
std::allocator<int*> >" due to the different spacing.
Change-Id: Ib4a57a808551d0b4cd527021f7d19772378e405f
Reviewed-by: Christian Stenger <christian.stenger@theqtcompany.com>
Seems it wasn't actually required at all
Change-Id: I4c622057b7fd85a64f99368eb07b502aee524955
Task-number: QTCREATORBUG-14217
Reviewed-by: Kai Koehne <kai.koehne@theqtcompany.com>
If you insert events with equal start times in a way that the model can
not reorder them by their end times the parenting follows the indexes,
not the end times. This has caused confusion multiple times and it's
good to test it.
Change-Id: I3e9ba212ff377dc2afa2a2893637bb55edebafea
Reviewed-by: Kai Koehne <kai.koehne@theqtcompany.com>
Failures were due to version specific reordering. This is still
checked in cases h1, h7, h8. Simplify maintanance by removing the
check in the other cases.
Change-Id: I481672e693f6370c75bc4739d6cca937b9f027ea
Reviewed-by: Christian Stenger <christian.stenger@theqtcompany.com>
Take new QPolygon{,F} dumpers into account, also make QUrlPrivate
field accessible,
Change-Id: I2b5c122895c3ee389ba939eaeffcd613fa206009
Reviewed-by: Christian Stenger <christian.stenger@theqtcompany.com>
Fixes crash in tst_generic_highlighter
Change-Id: I186cf47990e80ee6902c0812e6d6e3a260dfeec5
Reviewed-by: Christian Stenger <christian.stenger@theqtcompany.com>
Warnings from the scene graph renderer are most likely things we want
to catch when testing, such as failure to compile the shaders.
Change-Id: I96bcf4a44d0986ca6f065cda71b242ebb355ce84
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
That's for a structure mimicking QMakeStepConfig, comparing
a.x == b.x && a.y == b.y ... with
std::tie(a.x, a.y, ...) == std::tie(b.x, b.y, ...)
Executive summary: gcc 4.9.1 optimizes the traditional == && == better.
This is solely due to the sequence of four bools in the struct.
Change-Id: I77cc8d685c6b7b7ead651cd1f44ff1b2e9b39f3d
Reviewed-by: Daniel Teske <daniel.teske@theqtcompany.com>
... and further unify GDB and LLDB code paths.
Change-Id: Id89f3804c53190c4888082891fd3c3c55eceac84
Reviewed-by: Christian Stenger <christian.stenger@theqtcompany.com>
Make the expected output copyable to the test.
Change-Id: I4fc61682f14d0de054a11d2ffa4a0f36ede0e500
Reviewed-by: Francois Ferrand <thetypz@gmail.com>
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@theqtcompany.com>
