You can get all exported names. That can the C++ names too for C++
types. If you only want the types name which are imported in a document
you can provide the source id for the document and get only that. If
there are different version exports you get them too.
Task-number: QDS-10274
Change-Id: Iefc0b68ebfd65998cf0481961f8c7643cdf9e5b9
Reviewed-by: Vikas Pachdha <vikas.pachdha@qt.io>
It is matching the document imports with the exported type names.
Task-number: DS-10265
Change-Id: I0623cd0b3d04fa1a8c05354a194f10f3d8948b21
Reviewed-by: Vikas Pachdha <vikas.pachdha@qt.io>
Actually the syntax was wrong.
You can use
SKIP_AUTOMOC foo.cpp
or
PROPERTIES SKIP_AUTOMOC ON
SKIP_AUTOGEN is skipping AUTOUIC and AUTORCC too.
Change-Id: Id4afdc2572fc287ae66368ba0fd70b3ce7fc04ba
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
Reviewed-by: Burak Hancerli <burak.hancerli@qt.io>
metaInfo.isProjectComponent() is signaling if a file component is
directly part of the project. metaInfo.isInProjectModule() otherwise
shows if a type is inside of the project as a qmldir module. Both cannot
be true. If a type is not part of either it is a system type.
Task-number: QDS-10251
Change-Id: Iae2270827a500ad6393e3751b3af276f9b030679
Reviewed-by: Thomas Hartmann <thomas.hartmann@qt.io>
Reviewed-by: Qt CI Patch Build Bot <ci_patchbuild_bot@qt.io>
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
There were multiple reasons why the tests were flacky. First
Utils::reverseCompare had a bug. Now
std::lexicographical_compare(first.rbegin(), first.rend(),
second.rbegin(), second.rend())
is used.
Second the check StorageCache::checkEntries was not const. So it would
change the vector which it was iterating. So the iterator could be an
dangling.
Fixes: QDS-10197
Change-Id: I84fca6a2b24e1cae9fb85a01b6208de7e58240df
Reviewed-by: Qt CI Patch Build Bot <ci_patchbuild_bot@qt.io>
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
If you link to the internal Sqlite library you link actually twice. That
leads to strange bugs because you have an uninitialized object file. The
static linking workaround is now moved to cmake.
Change-Id: I51d966a623a18486ce5870d898999f3ce139f092
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
The function is not used but triggers strange accesses to null pointer.
Task-number: QDS-10178
Change-Id: Iea51609f7b8956a98fe9391aef8edac899e14c03
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
There were problems with older google test versions which were found
instead. Now the google test version is always the same and there cannot
be "strange" compile bugs.
Change-Id: Ib3dc74d1abbe369fb37a4ee5616011d8e3696c01
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Eike Ziller <eike.ziller@qt.io>
... with recent (namespaced) Qt dev.
[...]tests/unit/unittest/3rdparty/googletest/googletest/include/gtest/gtest-matchers.h:184:14: error: invalid conversion from 'const char*' to 'ushort' {aka 'short unsigned int'} [-fpermissive]
In file included from /data/dev/qt-6-ns/qtbase/include/QtCore/qstring.h:1,
from /data/dev/qt-6-ns/qtbase/include/QtCore/QString:1,
from /data/dev/creator-out/src/libs/utils/smallstringview.h:9,
from /data/dev/creator-out/src/libs/utils/smallstringliteral.h:8,
from /data/dev/creator-out/src/libs/utils/smallstring.h:8,
from /data/dev/creator-out/src/libs/utils/smallstringvector.h:8,
from /data/dev/creator-out/src/libs/utils/smallstringio.h:6,
from /data/dev/creator-out/tests/unit/unittest/gtest-creator-printing.h:7,
from /data/dev/creator-out/tests/unit/unittest/googletest.h:17,
from /data/dev/creator-out/tests/unit/unittest/externaldependenciesmock.h:6,
from /data/dev/creator-out/tests/unit/unittest/modulescanner-test.cpp:4:
a/dev/qt-6-ns/qtbase/src/corelib/text/qstring.h:1006:61: note: initializing argument 2 of 'bool Ns::operator==(const Ns::QString&, Ns::QChar)'
[104/1218] Building CXX object src/plugins/coreplugin/CMakeFiles/Core.dir/locator/executefilter.cpp.o^C
Change-Id: Id8d4533b9e53d7aa81edc64a99df31dd1d6ac5bf
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Marco Bubke <marco.bubke@qt.io>
Instead of putting everything into one big target there are now
libraries which are used by the tests.
Change-Id: I15727924f76b5cb600b7e533e7a6b8701dba81e0
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Burak Hancerli <burak.hancerli@qt.io>
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
A majority prefers snake case for long names and because test names are
describtive they have to be long.
Change-Id: I898beb5c9f1b9ab7f7c285b6278454068744f5da
Reviewed-by: Burak Hancerli <burak.hancerli@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
If an id is present in a binding expression and the node with that id is
removed we remove the expression is removed too.
Task-number: QDS-10018
Change-Id: I6c33005edcbcafbff7dd82af47977b238068eba1
Reviewed-by: <github-actions-qt-creator@cristianadam.eu>
Reviewed-by: Thomas Hartmann <thomas.hartmann@qt.io>
Sometimes multiple nodes or properties has to be removed. In that case
there is also the possibilty that we want to bypass the model resource
management.
Task-number: QDS-9766
Change-Id: I6c3cb0f682a7579f23d72431f641e2f812e2c63c
Reviewed-by: <github-actions-qt-creator@cristianadam.eu>
Reviewed-by: Thomas Hartmann <thomas.hartmann@qt.io>
This patch also includes a small update to skip creating 'shaderTool'
object in the json container if it doesn't exist in the qmlproject
file.
Task-number: QDS-9969
Change-Id: I1b8639c5e0a57aa77cbd417b51d57c1f4910613a
Reviewed-by: <github-actions-qt-creator@cristianadam.eu>
Reviewed-by: Thomas Hartmann <thomas.hartmann@qt.io>
The model is now deleting and changing the dependencies by itself. That
has the advantage of consistency. So if a node is now removed all the
dependecies will be deleted changing in the same way.
Task-number: QDS-9766
Change-Id: I1edf9ce172e51b9f52b62612ed2bb7386bf5b221
Reviewed-by: Thomas Hartmann <thomas.hartmann@qt.io>
Reviewed-by: <github-actions-qt-creator@cristianadam.eu>
The database has to be locked because otherwise it is not thread save.
Because Sqlite connections are already called in different threads this
has to be enforced. The easiest way to lock a database is to put it in a
transaction.
It should be ok without locking so long the connection is only called
from one thread but I prefer to enforce it because I have already seen
code was preparing a statement not on initialization stage where it is
highly unlikely that there is more than one thread pre connection but in
the middle of the code. Because preparing is much more expensive than a
lock I prefer to enfore a lock here.
Change-Id: Id0b47f8d615a6697bb807392cafbe976bdd37233
Reviewed-by: Vikas Pachdha <vikas.pachdha@qt.io>
Reviewed-by: <github-actions-qt-creator@cristianadam.eu>
There are many callback which do not control the flow and want to be
called for all steps. For that case we assume that a callback without a
return value will continue always.
Change-Id: I706123f2c425ac8c937d4d60ab977dae5ea9b030
Reviewed-by: Vikas Pachdha <vikas.pachdha@qt.io>
