[cfe-dev] Info : Clang Static Analyzer Statistics

[Anna Zaks]

Umesh,

There are two options for statistic tracking in the analyzer:
 - debug.Stats checker (AnalyzerStatsChecker.cpp), which currently reports coverage in terms of reachable(analyzed) basic blocks in top level functions. This checker runs at the very end when analysis are complete.
 - -analyzer-stats option is based on LLVM's stats option (http://llvm.org/docs/ProgrammersManual.html#Statistic). It enables printing of various analyzer internal counters and timers, it's extremely easy to extend.

You are welcome to submit patches for stats which are missing.

Currently, the number of warnings issued is not reported as a statistic, but is reported as output of the analyzer run (Ex: "5 warnings generated").

Cheers,
Anna.

On Jun 2, 2012, at 6:21 AM, Umesh Kalappa wrote:

> Hi anna,
> thanks  for the reply and im looking for the statistics like time taken for anaylsis ,loc,no of issues found and code coverage for cumulative analysis...
> 
> thanks again
> umesh
> 
> On Jun 1, 2012 11:15 PM, "Anna Zaks" <ganna at apple.com> wrote:
> 
> On Jun 1, 2012, at 4:46 AM, Umesh Kalappa wrote:
> 
> > Hi Ted and All ,
> >
> > Have a question regrading the clang analyzer statistics that was echoed  by clang when we use the switch "-analyze-stats"
> >
> 
> > We see that the statistics was echoed for every module i.e for  each  transition unit or file ,  can we configure clang to echo the statistics to  whole project ???
> > Other thing can we have LOC information too in the statistics  ???
> >
> 
> "analyzer-stats" turns on internal analyzer statistics used for debugging, analyzer coverage/performance measurements. What statistics are you interested in?
> 
> > Thanks
> > ~Umesh
> >
> > _______________________________________________
> > 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/20120604/4792281b/attachment.html>