Review Request: KDevelop C++ Checks

[Aleix Pol Gonzalez]

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://git.reviewboard.kde.org/r/104888/#review18479
-----------------------------------------------------------


Just to wake up a bit the discussion, I've been trying to play with the m_astToType by disabling it and by changing from QMap and QHash.

There's not really an impact regarding memory consumption or execution time between the 3 implementations.

I'd really like to put some stuff from here in practice, so can anybody at least back me up a bit on this?

- Aleix Pol Gonzalez


On May 9, 2012, 12:50 a.m., Aleix Pol Gonzalez wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://git.reviewboard.kde.org/r/104888/
> -----------------------------------------------------------
> 
> (Updated May 9, 2012, 12:50 a.m.)
> 
> 
> Review request for KDevelop, Milian Wolff and David Nolden.
> 
> 
> Description
> -------
> 
> This is the second big merge from my thesis. If we get this all the extra plugins will be usable already.
> 
> It basically provides the needed information to perform some checks: builds the control flow graph and the data access when requested (this information isn't created if it's not requested, by design, see kdevplatform review request).
> 
> The only place that might be not optimal is the creation of the m_AstToType map. This has to be discussed probably I'm unsure if it would be fine to just use the DUChain for this to extract the type from a cursor. I think the rest should be fine.
> 
> For more information, see: https://git.reviewboard.kde.org/r/103071/
> 
> 
> Diffs
> -----
> 
>   languages/cpp/cppduchain/CMakeLists.txt 5f27b8f 
>   languages/cpp/cppduchain/controlflowgraphbuilder.h PRE-CREATION 
>   languages/cpp/cppduchain/controlflowgraphbuilder.cpp PRE-CREATION 
>   languages/cpp/cppduchain/expressionvisitor.cpp 0ef76a6 
>   languages/cpp/cppduchain/tests/CMakeLists.txt 0bcd5e0 
>   languages/cpp/cppduchain/tests/codeanalysistest.h PRE-CREATION 
>   languages/cpp/cppduchain/tests/codeanalysistest.cpp PRE-CREATION 
>   languages/cpp/cppduchain/tests/testhelper.h 61a3de2 
>   languages/cpp/cppduchain/tests/testhelper.cpp 49e68a9 
>   languages/cpp/cppduchain/usebuilder.cpp f3f5926 
>   languages/cpp/cppduchain/usedecoratorvisitor.h PRE-CREATION 
>   languages/cpp/cppduchain/usedecoratorvisitor.cpp PRE-CREATION 
>   languages/cpp/cppparsejob.h debb1b0 
>   languages/cpp/cppparsejob.cpp 1e0a51b 
>   languages/cpp/parser/parsesession.h 1f69da7 
>   languages/cpp/parser/parsesession.cpp e1fcef5 
> 
> Diff: http://git.reviewboard.kde.org/r/104888/diff/
> 
> 
> Testing
> -------
> 
> There are quite some unit tests made, maybe not enough but that can be fixed :).
> 
> Also it can be used together with the kdevchecksrunner and kdevanalitzachecks to run some checks. Many projects have been run on top of those successfully (without crashes and with meaningful warnings). KDevelop for example.
> 
> 
> Thanks,
> 
> Aleix Pol Gonzalez
> 
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kdevelop-devel/attachments/20120904/7a6310d7/attachment.html>