Problem was with nested template type(in this case shared_ptr<T> is nested template type)
Task-number: QTCREATORBUG-6784
Change-Id: I8ea9203f2b1fbde73d9ac4e6c8cdeb2b6e0afcef
Reviewed-by: David Schulz <david.schulz@digia.com>
Problem was with cyclic recurrence.
To solve it we need to check if derived class is different class than its base class.
Keep completion corrected.
Include some unit tests when base class has the same name as derived.
Task-number: QTCREATORBUG-7887
Change-Id: I7973c0b06e3b62d2da3d06048f4327d18a0b8011
Reviewed-by: hjk <qthjk@ovi.com>
Change-Id: Ie49bdf576a6d3543aef6df133b27c8827775c935
Rationale:
a) It is conceptually cleaner to do the changes to the model
in between calls to beginResetModel() and endResetModel,
respectively.
b) The function is deprecated in Qt 5, for exactly that reason.
Reviewed-by: hjk <qthjk@ovi.com>
This time in the 'new' expression. Changed it to make
new C(1, abc...) and new C{1, abc}
work.
Change-Id: I7232798fd083b653ee04ef9ede386d6536133e16
Reviewed-by: hjk <qthjk@ovi.com>
Also drop the unused 'initializer' member from
RangeBasedForStatementAST.
Change-Id: I078ebbc85cafa643af4bfe62d698bf7de71360e4
Reviewed-by: hjk <qthjk@ovi.com>
This addresses the main memory leak revealed in QTCREATORBUG-7645.
The other leaks seem to have their origin in Qt.
Task-Number: QTCREATORBUG-7645.
Change-Id: I77f45449416c143b222ed5f5c905cba9674f95bb
Reviewed-by: Christian Kamm <kamm@incasoftware.de>
Make sure newly deprecated stuff is still available.
Change-Id: I9ebdfcd9a5ecee125a3c73f5f3254ae319d8b282
Reviewed-by: Joerg Bornemann <joerg.bornemann@digia.com>
It'll be reused as the initializer expression for declarators
that are followed by "( expression-list )".
Change-Id: I6c76a76641941874ef1ed21daa7b6e057c6d170f
Reviewed-by: hjk <qthjk@ovi.com>
When the template and base template are actually the same.
Task-number: QTCREATORBUG-7830
Change-Id: Ibf8ab5f5ee8da544ec768a078bd272500d2dc604
Reviewed-by: Christian Kamm <christian.d.kamm@nokia.com>
Even if "expand funcion-like macros" is unset we still
perform the expansion in the case it's already doing
so - when it originally started from an object-like macro.
Task-number: QTCREATORBUG-7712
Change-Id: Ie2a24de227f757d195146477d48246472082d28a
Reviewed-by: Orgad Shaneh <orgads@gmail.com>
Reviewed-by: hjk <qthjk@ovi.com>
There was an inconsistency, since the AST used in ResolveExpression
was not really the same previously computed. In the particular issue
below a crash could occur, for example, when using auto in a for
range loop.
Task-number: QTCREATORBUG-7828
Change-Id: I02958f434c3cf3b50609546003fc141674ee78d5
Reviewed-by: Eike Ziller <eike.ziller@nokia.com>
There was a fix for QTCREATORBUG-7730 in the case of nested
forward declarations in commit 74a458bca0.
However, it introduced regressions and actually didn't solve
the issue, since the behavior was hidden by another error fixed later.
The patch should properly fix the issue and the regression pointed
in QTCREATORBUG-7777.
Task-number: QTCREATORBUG-7730
Task-number: QTCREATORBUG-7777
Change-Id: I27397fefdc7cc9a60111761df1f76a01407886f7
Reviewed-by: Christian Kamm <christian.d.kamm@nokia.com>
Reviewed-by: Orgad Shaneh <orgads@gmail.com>
Previously we detected the duplicates by name and then processed
the missing ones, but this was not handling qualification properly.
Now we process the bases and only after lookup (making sure we
are talking about the *same* types) we add then if necessary.
Change-Id: Ic318b174d2174e24c24a4f2f1b612fbcb3f20491
Reviewed-by: Orgad Shaneh <orgads@gmail.com>
Reviewed-by: Roberto Raggi <roberto.raggi@nokia.com>
While a "real" C++ type rarely contains parantheses or other
characters that are "special" for a regexp, gdb happily
sprinkles in things like "(anonymous namespace)::".
Killing Creator while trying to beautify such a name
is inaccpetable.
This should be followed up by patches handling the known
problematic cases properly.
Change-Id: I8cc8509f1d7df0a8780876cdba556e1cf7ec4a95
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
Notice that a similar problem still exists for which we
need to fix the lexer when there's a C style commend which
ends with a backslash-newline.
Task-number: QTCREATORBUG-7713
Change-Id: I0f6d561703984f917fa5ed29de020ad0bdc5aaf0
Reviewed-by: David Schulz <david.schulz@nokia.com>
Reviewed-by: hjk <qthjk@ovi.com>
This fixes one of the issues mentioned in the report below.
THe other part will come in a separate patch.
Task-number: QTCREATORBUG-7730
Change-Id: I9f56a9bcec8a881dab3ab60f40c5b71f296466da
Reviewed-by: Roberto Raggi <roberto.raggi@nokia.com>
This fixes a variety of issues regarding class completion
when templates are used as base classes. The test cases
show examples.
Task-number: QTCREATORBUG-4357
Change-Id: I764d5ce817a78e1b19336e5beab758ca9e10f34b
Reviewed-by: Roberto Raggi <roberto.raggi@nokia.com>
