Example:
qregexparser.h(368): warning C4267: 'argument': conversion from 'size_t'
to 'int', possible loss of data

Change-Id: I025739d9dfa2d3ac4c254b59583575c3dae2f30e
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

It is necessary because when we reset the model we also reset its internal data.
Any previous requests (e.g. a header data request) will crash the application because
the data that is commning from remote doesn't fit anymore with the local data.

Change-Id: Iec1c4541d3440f8c4a560d9d81c47e14863bf5aa
Reviewed-by: Kevin Funk <kevin.funk@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

.gitattributes and .tag files were missing from qtremoteobjects

Task-number: QTBUG-58873
Change-Id: I5543e6dc6b57036d34813c5418615c4f44c9db2a
Reviewed-by: Akseli Salovaara <akseli.salovaara@qt.io>
Reviewed-by: Kevin Funk <kevin.funk@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

This patch adds the missing include statement where QRegExp is used.

Change-Id: Ia897c3e2c88dd4d5f7eceab21704162683db2a5c
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

Signature tests in qtremoteobjects seems to be flaky in windows so skip
those until fixed. There is no reason to delay qt5.9 alpha because of
flaky tests in new TP module

We cannot just blacklist some tests because problem is in test
compilation. That's why marking whole test set insignificant in windows

Task-number: QTBUG-58773
Change-Id: Ib264ff420d643a97e1592362f5ca73b0fe588d15
Reviewed-by: Liang Qi <liang.qi@qt.io>

This is an okay state, not an error.

Change-Id: I9e098ec50b8990d187770f6c17ba6546ff37e448
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

... and drop top_builddir, as it's unused.

Change-Id: I02d1c216b728a8ce8a07acf5d407388fa6d76838
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

Change-Id: I413ffc86f35119892401930f2627c812859a86ca
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

the CI obtains dependencies from the qt5 super repo nowadays, and it
makes no sense to put empty maps into it.

Change-Id: I41c49a5994c9852c73de07fcb5a55dbd12bca962
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

Change-Id: Iaed73151ad6777e861a7cc57674daad35a28afe3
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>

It should
1) Install to Framework when appropriate
2) Get copied in non-prefix/shadow builds
3) Only install once (avoid race conditions on copies)

Change-Id: I5f693873663d3609c8d153aaacbe01e11552a84d
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>

The build system will bootstrap automatically as needed (when cross-
building).

Task-number: QTPLAYGROUND-41
Change-Id: Ia4eddcfb7688a46e7f47660771a4cf9722c85136
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>

int/enum compare was not compiling on some systems

Change-Id: Ib8508d1d0502f80f84c1713c5a6b79cf5b8d78f8
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>

Two more instances of Q_DECL_OVERRIDE added

Change-Id: Iecc14df5a913d89500898a5ee2eafd1a99d73d09
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Michael Brasser <michael.brasser@live.com>

Change-Id: I980bcba3ff59aba16a52f137c469112d1942786c
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Andreas Holzammer <andreas.holzammer@kdab.com>

The Q_DECL_OVERRIDE should make OSX happy, and it looks like MSVC2013
doesn't like the Q_STATIC_ASSERT_X macro.

Change-Id: I2bd220159e46dc6c07c23b534eaa17cdbdb3c73b
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>

Commit d908ba2765cc7aef05832c8c442df6c1bf921820 in qtbase.git removes
the qt_check_for_QOBJECT_macro(...) function which we were relying on

Change-Id: I6da5bc650974197fdc580b35b10aa8e548a3d533
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>

Rename to QRegexParser, it isn't specific to .rep files, only the
parser.g file is specific to .rep files.

Change-Id: Ie74c8c996a38ea60cb2e1328e45178eb0582cdc4
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Michael Brasser <michael.brasser@live.com>

Change-Id: I3cf39180b08b80ea65c46e7b6a9e81f3aa79d1af
Reviewed-by: Kevin Funk <kevin.funk@kdab.com>
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>

It doesn't seem to like concatenating strings in a static assert.

Change-Id: I541ec7dad362f526d1f95ef21afddcc71c7aa2b1
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>

Change-Id: I09710e06ec674d282b4298a7f881a7ff229f924e
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Kevin Funk <kevin.funk@kdab.com>

Fixes the issue where code was protecting for
QRemoteObjectDynamicReplica to have a unique staticMetaObject.

Coverity-Id: 172024
Coverity-Id: 172029
Change-Id: I387f711f8d03a251217fef3712d7c2ed41544f2a
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Kevin Funk <kevin.funk@kdab.com>
Reviewed-by: Michael Brasser <michael.brasser@live.com>

Change-Id: I9f02c81114c913fd9c1056cede12ec8a344683f7
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

Make sure clients get their slot invocations ACK'd before shutting down
the server

Change-Id: I24ed944cad0af2e4af936cd746523f222e48cb89
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

This makes repc's behavior more consistent with other generators. If
command-line arguments are correct, input file exists, but the input is
malformed => remove output file. Before this patch, repc generated an
empty file.

On a consecutive make run, this empty file would be considered a valid
output / up-to-date and repc would not be re-run. This patch fixes this,
and makes sure repc is re-run each time, until the input is valid.

Change-Id: I17b20b6f0fb5a7d5b632077ae885540a90049afb
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

With this .rep file:
class MyClass
{
    SIGNAL(foo())
}

We got the following error:
.moc/../rep_classwithsignalonlytest_merged.h:153:11: error: class
‘MyClassSourceAPI<ObjectType>’ does not have any field named
‘methodArgTypes’
         , methodArgTypes{}

Fix this by not generating the code for the member initializers at all. The
members are initialized within the ctor anyway

Regression introduced by fc881f28



Change-Id: I8141aa799e02408a6484a13a28064813a2dc5007
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

Make sure QtRemoteObjects does not crash if we pass invalid urls to the
QRemoteObjectNode/QRemoteObjectHost constructor

Change-Id: I6db749ac7901f989cfaa73084d88fbd763ad9178
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

This includes addressing an issue where push<Prop> slots were not
defined correctly in the SourceApiMap.

Change-Id: I3791051f26c59d2b2cd1431cf166913b73d5fcba
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Michael Brasser <michael.brasser@live.com>

The coverity scan pointed to a bug in SourceApi where the intent was for
the static parameter checks (based on the static QObject::connect eval)
to return the list of argument types from the lookup.  It didn't before,
now it does.

Updated repc so autogenerated code includes the same changes.

Coverity-Id: 172053
Coverity-Id: 172054
Coverity-Id: 172060
Change-Id: I8e5ef9df028d5895153cdb2a97636218f64fb618
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Michael Brasser <michael.brasser@live.com>

Change-Id: Ifc97eb534b7ddd7dcc29df67846cb0afb782daff
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Kevin Funk <kevin.funk@kdab.com>
Reviewed-by: Michael Brasser <michael.brasser@live.com>

Addresses coverity scan warnings (Uninitialized scalar field)

Coverity-Id: 172058
Coverity-Id: 172059
Coverity-Id: 172204
Change-Id: Ief40a5a14fb7c64fb4e6776de24ef44e02d218be
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Michael Brasser <michael.brasser@live.com>

Also fixes coverity issue

Coverity-Id: 172061
Change-Id: Ic5d14efd878103fc105b7130d84aa068a85394e8
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Michael Brasser <michael.brasser@live.com>

Changes to QtRO properties by a Replica are inherently asynchronous.
The change needs to be propagated to the Source and then sent back to
all Replica objects.  This is counter-intuitive in many circumstances
(such as QML), where a call to a setter, followed by a call to the
corresponding getter is expected to return the "set" value.

With this change, QtRO will now create an assignProp() SLOT on both the
Source and Replica generated code.  The (virtual) Source method with
default to calling the setter.  The setter will not be available on the
Replica, making Source and Replica QObjects asymmetrical.

Change-Id: Ia1b1d12f59b3b6ebd7dea9ba566f76c2adf904e2
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Kevin Funk <kevin.funk@kdab.com>
Reviewed-by: Michael Brasser <michael.brasser@live.com>

- Add "we mean it" warning to private headers
- Properly include module headers in public headers

Change-Id: I3feded431f890fd451c07920edb102ae902a0fb9
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

Change-Id: Ife751435cf87778eda5f91dd786972ea3aad05c1
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

Change-Id: Iec2089a799bde83aa37d97a7bb139864b1f407ba
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Michael Brasser <michael.brasser@live.com>

Make sure QRemoteObjectReplica::state is populated properly when the
QRemoteObjectReplica::initialized() signal is emitted. Prior to this patch one
could end up in a state where initialized() was emitted with a
QRemoteObjectReplica::state == 0 (Uninitialized).

Change-Id: I1b55bad175b190706a0abc9e8253747cad1fa187
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

While being generated code it's still useful to make it human-readable
(since it's public API consumed by the QtRO user).

Changes:
- Move Q_PROPERTY decls to top
- Move enum decls + Q_ENUM{,s} decls to top
- Add newlines where appropriate

Change-Id: I2ecd597676c4feb6838ad40ec38969fbc30dfc8b
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>

They're not intended to be used from the outside (replica is
instantiated through a node).

Change-Id: I1fe8de5cd04c662d406800b6b6f62a56f57250cd
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>

Replicas already provide a mechanism to show a default value before the
connection to the Source is established.  This enhancement allows the
.rep PROP definition to include a 'PERSISTED' flag as well, which saves
the property value when the Replica is stopped and restores the value
when it is started again.

Because a shared dll (QtRO) should not know or care where and how values
are persisted, a persistence class needs to be set for the owner Node
for this feature to work.

Change-Id: I5718b024fd268c95d1add2b1be49e6864127782e
Reviewed-by: Continuous Integration (KDAB) <build@kdab.com>
Reviewed-by: Michael Brasser <michael.brasser@live.com>