The parent of instantiation of nested class of template class
should be the instantiation of enclosing template class.

To prevent the infinite loop for case with local typedef of enclosing
template we should not change a parent of typedefed instatiation of
enclosing template. Example:
template <typename T>
struct Enclosing
{
  typedef Enclosing<T> EnclosingT;// first case
  struct Nested
  {
    typedef Enclosing<T> EnclosingT;// second case
  };
};

Task-number: QTCREATORBUG-11752
Task-number: QTCREATORBUG-11999
Change-Id: Iadd7b5ef73ee0c4881f59c9dabfe03339f55827b
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>

Fix code completion for using pointer in template specialization and
initialization.
Example:
template <typename T>
struct S {};

template <typename T>
struct S<T*> { T* t; };

struct Foo { int foo; };

int main()
{
  S<Foo*> s;
  s.t-> //no code completion
  return 0;
}

Task-number: QTCREATORBUG-12638
Change-Id: Idcd461806a22f08b76236f2db6346f157b12f5d3
Reviewed-by: Orgad Shaneh <orgads@gmail.com>

Same patter. Plain is fully merged now, so remove the files, too.

Change-Id: Id8c0ba5689ad9980a0db3580cb9833344fd911f3
Reviewed-by: Tobias Hunger <tobias.hunger@digia.com>
Reviewed-by: Christian Stenger <christian.stenger@digia.com>

Fix code completion for case:
struct Foo
{
    int bar;
};

template <typename T>
struct Base
{
    T t;
};

template <typename T1 = Foo>
struct Derived : Base<T1>
{};

int main()
{
    Derived<> foo;
    foo.t.// no code completion
    return 0;
}

Task-number: QTCREATORBUG-12606
Change-Id: Iadf2fae172739d0a5844c6b437fd2686616e64e7
Reviewed-by: Erik Verbruggen <erik.verbruggen@digia.com>

Fix code completion in case:
struct Foo
{
  int bar;
};

template <typename T1 = Foo>
struct Derived : T1 { };

int main()
{
  Derived<> foo;
  foo. // members from Foo are not proposed
  return 0;
}

Task-number: QTCREATORBUG-12605
Change-Id: Ibe35c7b9a161e789057a4518c72390ac52489a3e
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>

Instead of having two lists of paths, now only one list is used where
both include paths and framework paths can be mixed. This reflects the
way the compiler is invoked, and retains the (correct) search order.

Task-number: QTCREATORBUG-11599
Change-Id: I373953e3e305df5b7a0d10920e12d146584adf9f
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>

Change-Id: I6db51aeba328dbb69ec59082a70be7bc2d95699b
Reviewed-by: Erik Verbruggen <erik.verbruggen@digia.com>

Task-number: QTCREATORBUG-12345
Change-Id: Ib2316ebdc81393b38185b9cb659fb85b78753e7b
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>
Reviewed-by: Eike Ziller <eike.ziller@digia.com>

This reverts commit 5d76b433 since it
might lead to an infinite loop.

Task-number: QTCREATORBUG-11999
Change-Id: I12c7f3afc4aa53a83ddc2bf2d33eb76c29b747ee
Reviewed-by: Eike Ziller <eike.ziller@digia.com>
Reviewed-by: Orgad Shaneh <orgads@gmail.com>

...where applicable.

Change-Id: Ie42e1012598adf124e5f66b6a53eda01724dfb71
Reviewed-by: Przemyslaw Gorszkowski <pgorszkowski@gmail.com>
Reviewed-by: Orgad Shaneh <orgads@gmail.com>

The parent of instantiation of nested class of template class
should be the instantiation of enclosing template class.

Task-number: QTCREATORBUG-11752
Change-Id: I8875c738848b85731f5ea88a60da190b4b0ad18f
Reviewed-by: Orgad Shaneh <orgads@gmail.com>
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>
Reviewed-by: Eike Ziller <eike.ziller@digia.com>

Fix a case:
struct A
{
  int a;
};

typedef A B;
typedef B* Bptr;
Bptr b;
b->

Task-number: QTCREATORBUG-10021
Change-Id: Ibf5ca801dbdb72744416924c0b8fc03daad5948a
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>

The case when anonymous class is inside function. Fixed:
* highlighting
* completion

Task-number: QTCREATORBUG-11711
Change-Id: Ic8fc5fdfb1aed62a74bf148ab7ed449d08214dda
Reviewed-by: Orgad Shaneh <orgads@gmail.com>
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>
Reviewed-by: Erik Verbruggen <erik.verbruggen@digia.com>

Case when nested class declaration contains object name for this class.
Example:
void fun()
{
    struct S
    {
        int i;
    } s;
    s.i;
}

Fixes:
* highlighting
* completion
* tests

Task-number: QTCREATORBUG-11710
Change-Id: I32e234f57655c388a87a199edc8be750d7bf823f
Reviewed-by: Orgad Shaneh <orgads@gmail.com>
Reviewed-by: Przemyslaw Gorszkowski <pgorszkowski@gmail.com>
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>

The binding in a block should not be used as an instantiationOrigin,
because it will be deleted if it does not contain any blocks or
namespaces or classes or enums.
Instead of that we should look for first enclosing class or namespace.

Task-number: QTCREATORBUG-11424
Change-Id: I0fc6e935495478f71372b0fe6f611887f45b2eda
Reviewed-by: Erik Verbruggen <erik.verbruggen@digia.com>

A member of nested anonymous class should be visible as a member of
enclosing class(if there is no declaration of this nested anonymous
class).

Fix:
* marking
* find usage
* follow symbol
* completion

Task-number: QTCREATORBUG-10876
Task-number: QTCREATORBUG-11170
Change-Id: If5b4d198e9075f2a8aa899ae59190f2c05f7b1ff
Reviewed-by: Orgad Shaneh <orgads@gmail.com>
Reviewed-by: Przemyslaw Gorszkowski <pgorszkowski@gmail.com>
Reviewed-by: Erik Verbruggen <erik.verbruggen@digia.com>

Discarded during merge from 3.0

Task-number: QTCREATORBUG-5456
Change-Id: I0bb4756e3cdf3c87a4c2b0fbfe6953faaa5e1c52
Reviewed-by: Przemyslaw Gorszkowski <pgorszkowski@gmail.com>
Reviewed-by: Erik Verbruggen <erik.verbruggen@digia.com>

Task-number: QTCREATORBUG-5456

Change-Id: I0bb4756e3cdf3c87a4c2b0fbfe6953faaa5e1c52
Reviewed-by: André Hartmann <aha_1980@gmx.de>
Reviewed-by: Erik Verbruggen <erik.verbruggen@digia.com>

Change-Id: Ib5423fdd064e4546f848c0b640b0ed0514c26d3a
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@digia.com>
Reviewed-by: Kai Koehne <kai.koehne@digia.com>

QVERIFY/QCOMPARE are meant to be called in the test function so that on
failure they just can "return" and thus skip subsequent code. Since we
use reusable test code in the test functions (the *TestCase classes), we
need to ensure that on failure no further test code is executed.

This mostly inlines the run function of the test classes into the
constructor.

Change-Id: I320ee032bdde0174ddfe3fdf3f9e18e19abf1d7f
Reviewed-by: Erik Verbruggen <erik.verbruggen@digia.com>

* Unify class names
* Fix coding style

Change-Id: I7498192ba9e6b9fc0e97d3d4f0dbb30f1853a0c9
Reviewed-by: Erik Verbruggen <erik.verbruggen@digia.com>

Move common functionality of the 12 test classes into base classes.

Change-Id: If64d3cec876807ac6f991151189860a99b8ff4ca
Reviewed-by: Erik Verbruggen <erik.verbruggen@digia.com>

Change-Id: If088533d0b51d1750e3273d7ab732dffd110ec85
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>

The bug was fixed long time ago, the test is missing a semicolon

Change-Id: I0236048d2da059fe208591469fb69541705463cf
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>

Comma is used to separate tests on command-line

Change-Id: I785d2137d54b0e1e4edbbd940042dc926719dcfc
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>

Change-Id: I9a5bf4705da5ca88776dec70057da2c64e4c1494
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>

Mostly done using the following Ruby script:

Fname = 'src/plugins/cpptools/cppcompletion_test.cpp'
s = File.read(Fname)
mod = s.gsub(/(?<declaration>void CppToolsPlugin::test_(?<test_name>[^(]+)\(\)\n\{)(?<body>.*?\n)\}/m) { |func|
  match = $~
  declaration = match['declaration']
  body = match['body']
  test_name = match['test_name'].sub('_data', '')
  final = body.gsub(/(?:completions.clear\(\);\s+)?(?: +)[^\n]* =\n(?<code>.*?);\s*(?<completions>(?:completions.append\(QLatin1String\("[^"]*"\)\);\s*)*)QTest::newRow\("case: (?<name>.+?)"\)\s*<< code << _\("(?<prefix>[^"]+)"\) << completions;/m) {
    m = $~
    res = "    QTest::newRow(\"#{test_name}: #{m['name']}\") << _(\n#{m['code'].rstrip}\n        ) << _(\"#{m['prefix']}\") << (QStringList()"
    m['completions'].scan(/completions.append\((.+)\);/) { |comp|
      res << "\n            << #{comp[0]}"
    }
    res + ');'
  }.gsub(/(?: +)[^\n]* =\n(?<code>.*?);\n\s*CompletionTestCase test\(.+?, "(?<prefix>.+?)"\);\s*QStringList expected;\s*(?<completions>(?:expected.append\(QLatin1String\("[^"]*"\)\);\s*)*)const QStringList completions = test.getCompletions\(\);\s*QCOMPARE\(completions, expected\);/m) {
    m = $~
    res = "    QTest::newRow(\"#{test_name}\") << _(\n#{m['code'].rstrip}\n        ) << _(\"#{m['prefix']}\") << (QStringList()"
    m['completions'].scan(/expected.append\((.+)\);/) { |comp|
      res << "\n            << #{comp[0]}"
    }
    res + ');'
  }.gsub(/(?: +)[^\n]* =\n(?<code>.*?);\n\s*CompletionTestCase test\(.+?(?:, (?<prefix>".+?"))?\);\s*(?:const )?QStringList completions = test.getCompletions\(\);\s*QCOMPARE\(completions.size\(\), \d+\);(?<completions>(?:\s*QVERIFY\(completions.contains\([^\n]+\);)*)\n/m) { |mm|
    m = $~
    res = "    QTest::newRow(\"#{test_name}\") << _(\n#{m['code'].rstrip}\n        ) << _(#{m['prefix']}) << (QStringList()"
    m['completions'].scan(/QVERIFY\(completions.contains\((.+?)\)\);/) { |comp|
      res << "\n            << #{comp[0]}"
    }
    res + ");\n"
  }.gsub(/(?: +)[^\n]* =\n(?<code>.*?);\n\s*CompletionTestCase test\(.+?(?:, (?<prefix>".+?"))?\);\s*bool replaceAccessOperator = false;\s*const QStringList completions = test.getCompletions\(\&replaceAccessOperator\);\s*QCOMPARE\(completions.size\(\), \d+\);(?<completions>(?:\s*QVERIFY\(completions.contains\([^\n]+\);)*)\s*QVERIFY\((?<replace>!?)[^)]*\);\n/m) { |mm|
    m = $~
    res = "    QTest::newRow(\"#{test_name}\") << _(\n#{m['code'].rstrip}\n        ) << _(#{m['prefix']}) << (QStringList()"
    m['completions'].scan(/QVERIFY\(completions.contains\((.+?)\)\);/) { |comp|
      res << "\n            << #{comp[0]}"
    }
    res + ")\n        << #{m['replace'].empty?};\n"
  }
  if final == body or final['QTest::addColumn']
    declaration + final + "}"
  else
    final
  end
}.gsub(/QTest::newRow\("([^"]+)"/) { |m|
  name = $1
  if name.size > 73
    space = name[0..73].rindex(/[ _]/)
    "QTest::newRow(\"#{name[0..space]}\"\n                  \"#{name[space+1..-1]}\""
  else
    m
  end
}.gsub(/\s+QTest::newRow/, "\n\n    QTest::newRow")
if mod != s
    File.open(Fname, 'wt').write(mod)
end

Change-Id: Id6bfb03cdf31ac27b36028fcdc861c340a5398f4
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>

For reasonable diff on following commit

Change-Id: I97e121355279f9166cea2f500cf2d4df5fafaa32
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>

Change-Id: I409a00b869a95dff2d69c9744da75c324c82776a
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>

Change-Id: I3c5f42de464a5607f1caa6a48d881c5a4a113ec3
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>

Change-Id: I2526daeaf54f7f8d4b808cf556cff9242adaa391
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>

Doesn't really make sense to have the additional IDocument *editorDocument()
method.

Change-Id: I0a7420eb1afaa76f63c3f7e9c4b373acf624ffb9
Reviewed-by: Eike Ziller <eike.ziller@digia.com>

Example:
namespace NS { struct S { int member; void fun(); }; }
using NS::S;
void S::fun()
{
mem// ctrl+space
}

Task-number: QTCREATORBUG-9037
Change-Id: I5a568be1b5c44deb02caa04996167a88a9c5d4d7
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>

This reverts commit 5be56c07 except for its tests. While it fixed the
problem described in the bug report, it also introduced regressions.

For now, we use a hard limit.

Change-Id: I19a85be454e3d7969371707e3eb1c3d42a16ecee
Reviewed-by: Przemyslaw Gorszkowski <pgorszkowski@gmail.com>
Reviewed-by: Erik Verbruggen <erik.verbruggen@digia.com>

This is the first phase of fixing bug QTCREATORBUG-10320.
This change resolves typedefs of template parameters(and resolves
problem with endless loop).

The next step will be matching appropriate template specialization
(this is needed to solve problem with missing code completion).

Missing matching: template specialization with the same parameters,
e.g.:
template <class T1, class T2, class T3>
class T
{
};

template <class T1, class T2>
class T<T1, T2, T2>
{
};

Task-number: QTCREATORBUG-10320
Change-Id: Icb6b539c021b2a67a66db9011a2e627f7d96526b
Reviewed-by: Przemyslaw Gorszkowski <pgorszkowski@gmail.com>
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>

Deleting an editor while currentEditorChanged is apparently not a good idea.
Use a queued connection instead.

This reverts commits 243a6259 and
d9602ca5.

Task-number: QTCREATORBUG-10190
Task-number: QTCREATORBUG-10235
Change-Id: Iccbee25fb77714963faa6f7184d1f5f53ed348a6
Reviewed-by: Eike Ziller <eike.ziller@digia.com>
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>
Reviewed-by: Jarek Kobus <jaroslaw.kobus@digia.com>

Deleting an editor while currentEditorChanged is apparently not a good
idea.
Use a queued connection instead.

This reverts commits 243a6259,
447c4ed3,
d9602ca5 and
31034870.

Change-Id: Iccbee25fb77714963faa6f7184d1f5f53ed348a6
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>
Reviewed-by: Eike Ziller <eike.ziller@digia.com>

Starting with commit 243a6259 EditorManager::closeEditors() does not
delete the editors anymore, but calls deleteLater(). When the call
returns, all kind of checks fail since the editor widget destructor was
not yet called.

Fixed by providing and calling Core::Tests::closeAndDeleteEditors(),
which explicitly sends QEvent::DeferredDelete to the widget.

Change-Id: I3287abbe74ccf7d7a18b997a847701c8be17cb55
Reviewed-by: Orgad Shaneh <orgads@gmail.com>
Reviewed-by: Eike Ziller <eike.ziller@digia.com>
Reviewed-by: Erik Verbruggen <erik.verbruggen@digia.com>

Since fuzzy completions are allowed, the lexicographically first proposal
is not necessarily most relevant. The patch modifies sorting of proposals
so that the exact match and continuations go first, and fuzzy completions
follow.

Moreover, being a continuation seem to be a more important characteristic
of a proposal, than being it a function argument or keyword etc. That's why
the check for continuation is placed before the check for order.

Task-number: QTCREATORBUG-8737
Task-number: QTCREATORBUG-9236
Change-Id: I89aae9d2ce6bfa59af7c2f75e6f3af00212008ca
Reviewed-by: André Hartmann <aha_1980@gmx.de>
Reviewed-by: Alexey Zhondin <lexxmark.dev@gmail.com>
Reviewed-by: Nikolai Kosjar <nikolai.kosjar@digia.com>

Change-Id: I3255a0150cd9a730336456c5a9f986eb74fefbff
Reviewed-by: Tobias Hunger <tobias.hunger@digia.com>