kdevelop:agateau-kdevelop.git
7 years agoset version to 4.0.2 4.0 v4.0.2
Milian Wolff [Thu, 2 Sep 2010 12:23:28 +0000 (14:23 +0200)]
set version to 4.0.2

7 years agodon't crash when renaming a file in custom make manager
Milian Wolff [Tue, 31 Aug 2010 16:25:24 +0000 (18:25 +0200)]
don't crash when renaming a file in custom make manager

BUG: 233824

7 years agoHelp people find the movingranges branch.
Andreas Pakulat [Thu, 19 Aug 2010 17:23:31 +0000 (19:23 +0200)]
Help people find the movingranges branch.

Provide the url for the devel-kdevelop clone which has the movingranges
branch so people can actually use it.

7 years agofix: don't skip comments inside strings inside macros
Milian Wolff [Tue, 17 Aug 2010 13:42:44 +0000 (15:42 +0200)]
fix: don't skip comments inside strings inside macros

7 years agoReally error out with too-new-kdelbis version
Andreas Pakulat [Tue, 17 Aug 2010 07:48:38 +0000 (09:48 +0200)]
Really error out with too-new-kdelbis version

Wrong argument to message()

7 years agoFix my blind max-version check.
Andreas Pakulat [Tue, 17 Aug 2010 05:42:47 +0000 (07:42 +0200)]
Fix my blind max-version check.

Somehow I put in the wrong minor version, sorry.

7 years agoMerge branch '4.0' of gitorious.org:kdevelop/kdevelop into 4.0
Andreas Pakulat [Mon, 16 Aug 2010 13:46:03 +0000 (15:46 +0200)]
Merge branch '4.0' of gitorious.org:kdevelop/kdevelop into 4.0

7 years agoError out if kdelibs is too new
Andreas Pakulat [Mon, 16 Aug 2010 13:44:50 +0000 (15:44 +0200)]
Error out if kdelibs is too new

Tell users of the git sources that KDevPlatform won't work with kdelibs
4.6.64 and later as there the smart-ranges have been removed. At some
point master can lift this requirement again, once the movingranges
branch has been merged, but 1.0 will never work with such kdelibs

7 years agofix compile... - sorry
Milian Wolff [Mon, 9 Aug 2010 12:13:00 +0000 (14:13 +0200)]
fix compile... - sorry

7 years agodon't offer ctors of indirect parent classes for overriding, add test
Milian Wolff [Mon, 9 Aug 2010 12:10:30 +0000 (14:10 +0200)]
don't offer ctors of indirect parent classes for overriding, add test

7 years agoFix maintainership info
Andreas Pakulat [Mon, 2 Aug 2010 18:58:00 +0000 (20:58 +0200)]
Fix maintainership info

I've resigned from maintainership, so adjust that. Aleix/Milian are now
co-maintaining.

7 years agostill expext fail, as this is not fixed yet
Milian Wolff [Sat, 31 Jul 2010 13:29:49 +0000 (15:29 +0200)]
still expext fail, as this is not fixed yet

7 years agodon't filter NamespaceAlias declarations when showing types
Milian Wolff [Sat, 31 Jul 2010 13:16:24 +0000 (15:16 +0200)]
don't filter NamespaceAlias declarations when showing types

7 years agoNULL tests matched to kdevelop coding standard
Ciprian Ciubotariu [Sat, 31 Jul 2010 11:07:25 +0000 (14:07 +0300)]
NULL tests matched to kdevelop coding standard

7 years agoAdded tests for handling cyclic namespace aliases
Ciprian Ciubotariu [Fri, 30 Jul 2010 17:16:41 +0000 (20:16 +0300)]
Added tests for handling cyclic namespace aliases

7 years agoDuchain tests for global namespace aliases
Ciprian Ciubotariu [Fri, 30 Jul 2010 00:34:45 +0000 (03:34 +0300)]
Duchain tests for global namespace aliases

7 years agoTest for code completion with namespace aliases
Ciprian Ciubotariu [Fri, 30 Jul 2010 00:33:31 +0000 (03:33 +0300)]
Test for code completion with namespace aliases

7 years agoFixed completion through global namespace aliases
Ciprian Ciubotariu [Fri, 30 Jul 2010 00:31:09 +0000 (03:31 +0300)]
Fixed completion through global namespace aliases

This is about the following scenarios:
namespace A = B;
using namespace A; // names in B get imported
A:: // names from B are presented

7 years agomake sure we actualy get a TemplateDeclaration before accessing it
Milian Wolff [Thu, 29 Jul 2010 15:08:30 +0000 (17:08 +0200)]
make sure we actualy get a TemplateDeclaration before accessing it

BUG: 246165

7 years agomark overridden explicit ctors also as explicit
Milian Wolff [Wed, 28 Jul 2010 20:40:10 +0000 (22:40 +0200)]
mark overridden explicit ctors also as explicit

7 years agoSVN_SILENT made messages (.desktop file)
Script Kiddy [Wed, 28 Jul 2010 02:17:51 +0000 (04:17 +0200)]
SVN_SILENT made messages (.desktop file)

7 years agocherry pick #c0aff55a17 from master: filter anonymous structs during code completion...
Milian Wolff [Sat, 24 Jul 2010 19:34:50 +0000 (21:34 +0200)]
cherry pick #c0aff55a17 from master: filter anonymous structs during code completion, add test

TODO: shows bug in .[indexed]Identifier().isEmpty() (i.e. that returns false) even though .toString().isEmpty() returns true...

Conflicts:

languages/cpp/tests/test_cppcodecompletion.cpp
languages/cpp/tests/test_cppcodecompletion.h

7 years agoSVN_SILENT made messages (.desktop file)
Script Kiddy [Wed, 21 Jul 2010 02:14:08 +0000 (04:14 +0200)]
SVN_SILENT made messages (.desktop file)

7 years agoSVN_SILENT made messages (.desktop file) v4.0.1
Script Kiddy [Fri, 16 Jul 2010 02:09:27 +0000 (04:09 +0200)]
SVN_SILENT made messages (.desktop file)

7 years agofix overflow in CMakeFunctionArgument::unescapeValue
Milian Wolff [Wed, 14 Jul 2010 14:26:51 +0000 (16:26 +0200)]
fix overflow in CMakeFunctionArgument::unescapeValue

BUG: 244581

7 years agoSVN_SILENT made messages (.desktop file)
Script Kiddy [Wed, 14 Jul 2010 02:11:29 +0000 (04:11 +0200)]
SVN_SILENT made messages (.desktop file)

7 years agoSVN_SILENT made messages (.desktop file)
Script Kiddy [Sat, 10 Jul 2010 02:03:24 +0000 (04:03 +0200)]
SVN_SILENT made messages (.desktop file)

7 years agoRemove some conflicting shortcuts from the katepart.
Mantia Andras [Thu, 8 Jul 2010 21:27:02 +0000 (00:27 +0300)]
Remove some conflicting shortcuts from the katepart.

7 years agofix: allow incr as identifier, it's not the ++ token after all
Milian Wolff [Thu, 8 Jul 2010 17:54:52 +0000 (19:54 +0200)]
fix: allow incr as identifier, it's not the ++ token after all

7 years agodon't crash on nullptr type in getPointerTarget
Milian Wolff [Wed, 7 Jul 2010 08:54:08 +0000 (10:54 +0200)]
don't crash on nullptr type in getPointerTarget

BUG: 243758

7 years agoMerge commit 'refs/merge-requests/18' of gitorious.org:kdevelop/kdevelop into merge...
Milian Wolff [Tue, 6 Jul 2010 21:53:13 +0000 (23:53 +0200)]
Merge commit 'refs/merge-requests/18' of gitorious.org:kdevelop/kdevelop into merge-request-18

7 years agofix shouldStartCompletion when writing a path in a #include, inserted is not only...
Milian Wolff [Tue, 6 Jul 2010 19:20:24 +0000 (21:20 +0200)]
fix shouldStartCompletion when writing a path in a #include, inserted is not only a single char

7 years agoSVN_SILENT made messages (.desktop file)
Script Kiddy [Sun, 4 Jul 2010 02:10:27 +0000 (04:10 +0200)]
SVN_SILENT made messages (.desktop file)

7 years agodon't crash when getContainerType does not find a declaration for the identified...
Milian Wolff [Thu, 1 Jul 2010 09:26:55 +0000 (11:26 +0200)]
don't crash when getContainerType does not find a declaration for the identified type in the given topcontext

7 years agoMerge remote branch 'origin/4.0' into 4.0
Andreas Pakulat [Thu, 1 Jul 2010 06:46:00 +0000 (08:46 +0200)]
Merge remote branch 'origin/4.0' into 4.0

7 years agoFix opening files with relative path.
Andreas Pakulat [Thu, 1 Jul 2010 06:42:42 +0000 (08:42 +0200)]
Fix opening files with relative path.

Apparently KUrl::isRelative doesn't work as we expect it to when
creating the url with a filename (not sure why), so simply use
QFileInfo instead.

7 years agoSVN_SILENT made messages (.desktop file)
Script Kiddy [Thu, 1 Jul 2010 02:07:39 +0000 (04:07 +0200)]
SVN_SILENT made messages (.desktop file)

7 years agoSet CMAKE_INSTALL_RPATH to lib-installdir
Andreas Pakulat [Wed, 30 Jun 2010 08:29:40 +0000 (10:29 +0200)]
Set CMAKE_INSTALL_RPATH to lib-installdir

This is necessary so the targets have rpath set even if installing into
a fresh empty prefix. If its not set the rpath is empty.

7 years agoMake sure kdevplatform is properly linked
Andreas Pakulat [Sun, 20 Jun 2010 09:00:22 +0000 (11:00 +0200)]
Make sure kdevplatform is properly linked

Setting RPATH is important so that installations outside of /usr still
work properly. Else users have to go fiddling with environment variables
or /etc/ls.so.conf

7 years agouse canonicalFilePath as absoluteFilePath might lead to problems when resolving relat...
Milian Wolff [Sun, 20 Jun 2010 19:09:03 +0000 (21:09 +0200)]
use canonicalFilePath as absoluteFilePath might lead to problems when resolving relative includes in symlinked directories

7 years agoFix another possible leak
Nicolás Alvarez [Sat, 19 Jun 2010 17:32:46 +0000 (14:32 -0300)]
Fix another possible leak

Peter Oberndorfer pointed out that my previous leak fix wasn't enough.

7 years agoFix memory leak in CMakeListsParser::readCMakeFile.
Nicolás Alvarez [Sat, 19 Jun 2010 15:40:31 +0000 (12:40 -0300)]
Fix memory leak in CMakeListsParser::readCMakeFile.

7 years agodon't advance over EOF, might lead to invalid tokens and crashes
Milian Wolff [Sat, 19 Jun 2010 10:51:57 +0000 (12:51 +0200)]
don't advance over EOF, might lead to invalid tokens and crashes

BUG: 242148

7 years agoPrevent calling a null pointer.
Aleix Pol [Sat, 12 Jun 2010 06:47:46 +0000 (08:47 +0200)]
Prevent calling a null pointer.

BUG: 240740

7 years agoSVN_SILENT made messages (.desktop file)
Script Kiddy [Thu, 10 Jun 2010 01:50:48 +0000 (03:50 +0200)]
SVN_SILENT made messages (.desktop file)

7 years agoAdd proper support for the Break command.
Aleix Pol [Mon, 7 Jun 2010 22:46:10 +0000 (00:46 +0200)]
Add proper support for the Break command.

7 years agoDon't try to open a non-existing file.
Andreas Pakulat [Mon, 7 Jun 2010 16:19:22 +0000 (18:19 +0200)]
Don't try to open a non-existing file.

the main.cpp is in the project directory not a src subdir.

7 years agoSVN_SILENT made messages (.desktop file)
Script Kiddy [Mon, 31 May 2010 02:02:41 +0000 (04:02 +0200)]
SVN_SILENT made messages (.desktop file)

7 years agomake __volatile__ a synonym for volatile
Milian Wolff [Tue, 25 May 2010 10:13:36 +0000 (12:13 +0200)]
make __volatile__ a synonym for volatile

BUG: 238772

7 years agodon't recurse into the same class in typeutils::getMemberFunctions
Milian Wolff [Thu, 20 May 2010 09:50:42 +0000 (11:50 +0200)]
don't recurse into the same class in typeutils::getMemberFunctions

7 years agoFind build directory even in root CMakeLists.txt
Esben Mose Hansen [Fri, 14 May 2010 14:34:02 +0000 (16:34 +0200)]
Find build directory even in root CMakeLists.txt

7 years agofix uses in argument lists of pure virtuals
Milian Wolff [Tue, 18 May 2010 14:31:12 +0000 (16:31 +0200)]
fix uses in argument lists of pure virtuals

the = 0; will show up as an Initializer in the AST while this is of course not applicable in a Class context.
fixes test

7 years agoUse config mode as fallback if no module is found, and set package_DIR in all cases...
Esben Mose Hansen [Sat, 15 May 2010 12:18:37 +0000 (14:18 +0200)]
Use config mode as fallback if no module is found, and set package_DIR in all cases where a module was not found

7 years agoTurn an assertion into a warning
David nolden [Sat, 15 May 2010 10:21:46 +0000 (12:21 +0200)]
Turn an assertion into a warning
BUG: 236411

7 years agoRemove support for building the whole extragear/sdk.
Niko Sams [Sat, 15 May 2010 05:14:54 +0000 (07:14 +0200)]
Remove support for building the whole extragear/sdk.

It's not possible anymore since the move to git

7 years agoRemove FindCppUnit.cmake now that the CppUnit plugin no longer exists
Kovarththanan Rajaratnam [Thu, 13 May 2010 05:26:04 +0000 (07:26 +0200)]
Remove FindCppUnit.cmake now that the CppUnit plugin no longer exists

7 years agoRemove FindCheck.cmake now that the Check plugin no longer exists
Kovarththanan Rajaratnam [Thu, 13 May 2010 05:35:39 +0000 (07:35 +0200)]
Remove FindCheck.cmake now that the Check plugin no longer exists

7 years agoIncrease version number
Andreas Pakulat [Wed, 12 May 2010 19:54:55 +0000 (21:54 +0200)]
Increase version number

This should always be set to the "next" version number after releasing
something. So do that now and increase for 4.0.1

7 years agoQuote variable expressions correctly.
Niko Sams [Sat, 8 May 2010 13:18:22 +0000 (15:18 +0200)]
Quote variable expressions correctly.

This fixes expressions containing spaces.

Thanks to Valentyn Pavliuchenko for the patch.
http://reviewboard.kde.org/r/3935
BUG: 235759

7 years agoset invalid range on final location of problem when include path resolver found more...
Milian Wolff [Thu, 6 May 2010 22:33:39 +0000 (00:33 +0200)]
set invalid range on final location of problem when include path resolver found more paths than the build manager

the 'did not return a build dir' error does the same

note: imo we should remove this problem I'm changing here alltogether though - it shows a limitation in the cmake
plugin that is only interesting to developers, not the user

7 years agoFix kdevelop app icon for windows.
Andreas Pakulat [Thu, 6 May 2010 15:29:08 +0000 (17:29 +0200)]
Fix kdevelop app icon for windows.

The split of main.cpp produced this, I overlooked the use of the SRCS
variable in kde4_add_app_icon.

Thanks to Krzysztof Nowicki for the patch.
BUG:236545

7 years agoUse a mutex to prevent setupStandardMacros() and setupStandardIncludePaths() from...
David nolden [Thu, 6 May 2010 10:12:15 +0000 (12:12 +0200)]
Use a mutex to prevent setupStandardMacros() and setupStandardIncludePaths() from being called multiple times at the same time, preventing a crash.
    BUG: 236434

7 years agoFix compile warnings related to deprecated hash_map
Bertjan Broeksema [Wed, 5 May 2010 09:49:55 +0000 (11:49 +0200)]
Fix compile warnings related to deprecated hash_map

g++ >= 4.3 has the unordered_map header/class as replacement for the now
deprecated hash_map header. This commit takes that in account to fix a
bunch of compile warnings related to this. I made sure this still works
for g++ < 4.3 and did *not* change anything related to MSVC builds. The
latter might (still) have a similar warning.

When LTS for g++ < 4.3 has ended, the include should be changed to
<unordered_map> and the equally named type should be used directly.

NOTE: The hash_map was profiled by David Noden to be better performing
than QHash. I didn't do additional profiling for unordered_map but
assume it performs equally as hash_map.

7 years agoAlso search for qmake-qt4
Andreas Pakulat [Tue, 4 May 2010 13:58:01 +0000 (15:58 +0200)]
Also search for qmake-qt4

Distributions providing qt3+qt4 usually rename qmake to include the version
number. This breaks the code trying to find the Qt docs. This change
fixes that by also taking qmake-qt4 into account.

BUG: 236315

7 years agoRevert r1121158 (fix linker error).
Andreas Pakulat [Sat, 1 May 2010 13:22:46 +0000 (13:22 -0000)]
Revert r1121158 (fix linker error).

If your linker cannot link the kdevelop binary you'll have to find a
different fix. All functions used from sessioncontroller.h are inline
and hence no linking is needed. I'm reverting this because linking
actually causes problems as it pulls in libkdevplatformduchain and that
one has global statics which must not be initialized in the kdevelop
binary but only once the kdevelop.bin is being loaded (thats why the
kdevelop.bin links against shell).

CCMAIL: brandon.ml@gmail.com

7 years agofix linker error
Carlo Segato [Fri, 30 Apr 2010 15:33:50 +0000 (15:33 -0000)]
fix linker error

7 years agoSVN_SILENT made messages (.desktop file) v4.0.0
Script Kiddy [Wed, 28 Apr 2010 06:30:29 +0000 (06:30 -0000)]
SVN_SILENT made messages (.desktop file)

7 years agoIncrease version number for final.
Andreas Pakulat [Tue, 27 Apr 2010 22:20:00 +0000 (22:20 -0000)]
Increase version number for final.

Preparation for the KDevelop4.0.0/KDevPlatform 1.0.0 release.

7 years agocleanup includes
Milian Wolff [Mon, 26 Apr 2010 21:24:24 +0000 (21:24 -0000)]
cleanup includes

7 years agoconvert the mimetype definition to the XDG format and remove the old KDE 3 definition
Fathi Boudra [Mon, 26 Apr 2010 09:48:58 +0000 (09:48 -0000)]
convert the mimetype definition to the XDG format and remove the old KDE 3 definition

7 years agoSVN_SILENT made messages (.desktop file)
Script Kiddy [Mon, 26 Apr 2010 06:26:20 +0000 (06:26 -0000)]
SVN_SILENT made messages (.desktop file)

7 years agoFix crash when attachMaybe is called a second time before the first -var-create call...
Niko Sams [Sun, 25 Apr 2010 20:00:34 +0000 (20:00 -0000)]
Fix crash when attachMaybe is called a second time before the first -var-create call returned

BUG: 219570

7 years agoRemove configuration for "Floating Toolbar" as it's implementation got
Niko Sams [Sun, 25 Apr 2010 18:31:41 +0000 (18:31 -0000)]
Remove configuration for "Floating Toolbar" as it's implementation got
somewhere lost.

7 years agoremove unused member
Niko Sams [Sun, 25 Apr 2010 18:31:39 +0000 (18:31 -0000)]
remove unused member

7 years agoHide broken option "Use separate terminal for I/O" for now.
Niko Sams [Sun, 25 Apr 2010 18:31:34 +0000 (18:31 -0000)]
Hide broken option "Use separate terminal for I/O" for now.

I can't fix this for 4.0 so just hide the option for now.
CCBUG: 235335

7 years agoremove expect_fail, this passes now
Milian Wolff [Sat, 24 Apr 2010 13:26:02 +0000 (13:26 -0000)]
remove expect_fail, this passes now

7 years agoadapt test to recent change in Q_PRIVATE_SLOT
Milian Wolff [Sat, 24 Apr 2010 13:26:01 +0000 (13:26 -0000)]
adapt test to recent change in Q_PRIVATE_SLOT

7 years ago- Add a maximum depth for recursive make calls
David Nolden [Sat, 24 Apr 2010 13:15:24 +0000 (13:15 -0000)]
- Add a maximum depth for recursive make calls
- Don't follow recursive make calls if the working directory didn't change
BUG: 232972

7 years agodon't assume ctor calls in class context, fixes method parsing with undefined type...
Milian Wolff [Sat, 24 Apr 2010 13:13:23 +0000 (13:13 -0000)]
don't assume ctor calls in class context, fixes method parsing with undefined type arguments

review: http://reviewboard.kde.org/r/3581/

7 years ago- Allow incomplete matching between the signal/slot specification.
David Nolden [Sat, 24 Apr 2010 13:02:09 +0000 (13:02 -0000)]
- Allow incomplete matching between the signal/slot specification.
- Fix the according test. The problem was not the fact that the signal has a default argument, rather it is _always_ possible to leave out signal arguments, when at least as many are specified as in the following reference to the slot.

7 years agoImprove our Q_PRIVATE_SLOT replacement macro so that:
David Nolden [Sat, 24 Apr 2010 12:33:16 +0000 (12:33 -0000)]
Improve our Q_PRIVATE_SLOT replacement macro so that:
- It has an inline implementation, and thus isn't shown in the implementation-helper
- The inline implementation contains a use of the "d" object, so that use-highlighting, refactoring, etc. work on it.

7 years agoDon't propose implementation-helpers for declarations with empty ranges, as those...
David Nolden [Sat, 24 Apr 2010 12:31:43 +0000 (12:31 -0000)]
Don't propose implementation-helpers for declarations with empty ranges, as those came from macros like Q_OBJECT.

7 years ago- Fill the ducontext member before calling "evaluateType" to prevent an assertion.
David Nolden [Sat, 24 Apr 2010 11:36:57 +0000 (11:36 -0000)]
- Fill the ducontext member before calling "evaluateType" to prevent an assertion.
- Add a more strict warning to make sure this is not forgotten.
Please, more testing next time, a simple run of kdevelop was enough to notice the problem.

7 years agotypeof-support with unit-testing
Jonathan Schmidt-Dominé [Fri, 23 Apr 2010 23:37:12 +0000 (23:37 -0000)]
typeof-support with unit-testing

7 years agotypeof works correctly, unit test will follow when I have fixed the environment...
Jonathan Schmidt-Dominé [Fri, 23 Apr 2010 12:42:23 +0000 (12:42 -0000)]
typeof works correctly, unit test will follow when I have fixed the environment...
TODO: Ignore optional typename-keywords, e.g. typedef typename __if_type::__type __type in type_traits.h

7 years agoFixed the typeof, ext/type_traits.h works without problems
Jonathan Schmidt-Dominé [Fri, 23 Apr 2010 12:09:20 +0000 (12:09 -0000)]
Fixed the typeof, ext/type_traits.h works without problems

7 years agoReduced the problems with standard-headers.
Jonathan Schmidt-Dominé [Fri, 23 Apr 2010 11:56:18 +0000 (11:56 -0000)]
Reduced the problems with standard-headers.
(__asm, __asm__, __typeof__, typeof)
TODO: It should be typeof(typeid | expression) and not typeof(typeid)

7 years agodon't offer implementation helper for signals, you never want to do that
Milian Wolff [Fri, 23 Apr 2010 11:14:05 +0000 (11:14 -0000)]
don't offer implementation helper for signals, you never want to do that

7 years agodon't skip static member functions when doing code completion on an object
Milian Wolff [Thu, 22 Apr 2010 14:45:09 +0000 (14:45 -0000)]
don't skip static member functions when doing code completion on an object

this really bugs me since it's perfectly fine CPP code, and is used especially often with e.g. QApplication app; app.exec(); and similar

add test, also check for the static access, where nothing gets filtered. Can someone elaborate on the why of this comment:
///@todo what NOT to show on static member choose? Actually we cannot hide all non-static functions, because of function-pointers

CCMAIL: david.nolden.kdevelop@art-master.de

7 years agoRevert "Integrate the foreground lock."
Andreas Pakulat [Mon, 19 Apr 2010 21:09:37 +0000 (21:09 -0000)]
Revert "Integrate the foreground lock."

This is too risky to add this late in the release cycle and additionally
shouldn't be added to 4.0 at all (only to 4.1).

7 years agoAdd missing i18n() calls to get all strings translatable.
Aleix Pol Gonzalez [Mon, 19 Apr 2010 00:21:22 +0000 (00:21 -0000)]
Add missing i18n() calls to get all strings translatable.

Thanks to Nicolás Alvarez for reporting

7 years agoAdapt to the kdevplatform changes. That fixes the bug discussed on the list that...
Aleix Pol Gonzalez [Sun, 18 Apr 2010 18:29:40 +0000 (18:29 -0000)]
Adapt to the kdevplatform changes. That fixes the bug discussed on the list that triggered a crash when renaming a file in many targets.

7 years agoReplace global static with call to qapp->currentthread.
Andreas Pakulat [Sun, 18 Apr 2010 16:53:17 +0000 (16:53 -0000)]
Replace global static with call to qapp->currentthread.

Provides the same logic, but doesn't rely on some global static being
initialized properly.

7 years ago- Add an object that releases the top-context on its destruction, so that it's also...
David Nolden [Sun, 18 Apr 2010 13:28:15 +0000 (13:28 -0000)]
- Add an object that releases the top-context on its destruction, so that it's also released when a test has failed.
- Use that object in the testLocalNamespaceAlias test, so that it doesn't make another test fail. Now all tests pass again.

7 years agoAdd a test for bug 207548
David Nolden [Sun, 18 Apr 2010 13:05:51 +0000 (13:05 -0000)]
Add a test for bug 207548
CCBUG: 207548

7 years agoMore precisely search the scopes for external class definitions.
David Nolden [Sun, 18 Apr 2010 12:30:52 +0000 (12:30 -0000)]
More precisely search the scopes for external class definitions.
Add test.
BUG: 204100

7 years agoMake the include-paths unique before computing a list of shown items.
David Nolden [Sun, 18 Apr 2010 12:04:05 +0000 (12:04 -0000)]
Make the include-paths unique before computing a list of shown items.
BUG: 202321

7 years agoProvide a stub implementation of __builtin_offsetof that allows correct use-building...
David Nolden [Sun, 18 Apr 2010 11:51:50 +0000 (11:51 -0000)]
Provide a stub implementation of __builtin_offsetof that allows correct use-building for the member.
BUG: 187046

7 years agoFix build
David Nolden [Sun, 18 Apr 2010 11:32:43 +0000 (11:32 -0000)]
Fix build

7 years agoAdd a zero-check after locking the duchain, to prevent a rare crash.
David Nolden [Sun, 18 Apr 2010 11:13:19 +0000 (11:13 -0000)]
Add a zero-check after locking the duchain, to prevent a rare crash.
BUG: 231790