On Sep 20, 2010, at 10:05 PM, Marcin Świderski wrote:

> Added comments.

Thanks Marcin.  The contents of the patch look good to me.

> 
> I thought about tests myself but don't really know how should we go about 
> them? I mean that for now destructors will probably be off by default. How to 
> print CFG using the driver? How to set up testing the output?

The CFG can be printed from the command line thusly:

$ clang -cc1 -analyze -cfg-dump <file.cpp>

we can then check the output using FileCheck.  A good analogous test is 
plist-output.m, which tests the plist output of the analyzer that is currently 
used for rendering paths in Xcode (and potentially other clients).

> 
> W dniu 21 września 2010 03:24 użytkownik Ted Kremenek <[email protected]> 
> napisał:
> Looks great!
> 
> Please add doxygen comments about CFGInitializer and CFGAutomaticObjDtor 
> respectively.  We should get in the habit of documenting such core classes in 
> the static analyzer.
> 
> We also should probably add a FileCheck-based test to test/Analysis that 
> tests CFG printing (particularly these new pieces), since it has broken in 
> the past.  Now that we're adding new CFGElements, now seems like a good time 
> to start testing this.
> 
> On Sep 20, 2010, at 4:42 PM, Marcin Świderski wrote:
> 
> > This patch:
> > - defines interfaces for CFGInitializer and CFGAutomaticObjDtor,
> > - adds support for above classes to print_elem function (renamed print_stmt)
> >
> > I've extended StmtPrinterHelper to print locations of VarDecls for 
> > destructors.
> >
> > Please aprove for commit,
> >
> > Marcin
> > <cfg-print-elem.patch>
> 
> 
> <cfg-print-elem.patch>

_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits

Reply via email to