[cfe-dev] GSOC Static Analyzer Proposal

[Adam Schnitzer]

Anna,

Thank you for the feedback. I am working on identifying checkers that would
be valuable, and can
be implemented in the existing static analyzer infrastructure. I will
report back with a draft of
planned deliverables and implementation plans.

I am also working on one of the simpler checkers to get some practice with
the analyzer.
The "Unary Plus with Unsigned" checker seemed like a good place to get
started, unless
you think there is one that might be better.

Adam

On Thu, Apr 11, 2013 at 1:30 PM, Anna Zaks <ganna at apple.com> wrote:

>
> On Apr 10, 2013, at 9:44 AM, Sean Silva <silvas at purdue.edu> wrote:
>
>
>
>
> On Tue, Apr 9, 2013 at 11:51 PM, Adam Schnitzer <adamschn at umich.edu>
> wrote:
>
>> Hello All,
>>
>> I am planning on proposing a project for Google Summer of Code this
>> summer, and would like to get your feedback before I write up a formal
>> proposal.
>>
>> I would like to work on improving support for C++ in the static analyzer.
>> Specifically, I think it would be valuable to improve the checkers for
>> undefined behavior including those already suggested<http://clang-analyzer.llvm.org/potential_checkers.html>
>> .
>>
>> Also, I think it would be helpful to extend the static analyzer to check
>> for stylistic violations. For example, projects like LLVM have suggestions
>> like, "Don't use else after a return". These warnings would often be noisy,
>> and project dependent, so it would be useful to make those options
>> configurable and suppressible.
>>
>> I am also interested in implementing several of the optimization
>> checkers. Specifically, it would be valuable to have warnings about postfix
>> increment and pass by value give an idea of how large the object being
>> copied is.
>>
>
>
> All these ideas sound valuable. I recommend initially focusing on just one
> of them though, and get to the others as time permits.
>
>
> +1
>
> It would be good to focus on one or several of checks and have a concrete
> plan of how they will be implemented. You can send the specifics to the
> list to get some feedback. Additional bonus would be to choose the checks
> that are not currently covered by other tools in clang/llvm family but this
> is not a requirement in my opinion.
>
> I would choose something from C++ undefined behavior or the optimization
> buckets you mention above. Try to aim for the checks that have a high
> payoff (in the number of possible bugs/bug severity) and are easier to
> implement in the static analyzer (do not require too much infrastructure
> support).
>
> Anna.
>
> -- Sean Silva
> _______________________________________________
> cfe-dev mailing list
> cfe-dev at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20130411/641a51f4/attachment.html>