Re: [csmith-dev] [extended-csmith] CODE COVERAGE ON GCC

[sameeran joshi <gsocsameeran@gmail.com>]

On 12/23/18, Yang Chen <chenyang@cs.utah.edu> wrote:
> Hi Sameeran,
>
>
> On 12/22/18 9:17 PM, Sameeran Joshi wrote:
>> As I had read it in your paper 'Finding And Understanding Bugs in C
>> Compiler' about the code coverage done for Csmith, Can you point out
>> the way you did the code coverage, any blogs written by you people or
>> some links in 'doing code coverage in gcc' ?
>>
>
> We used lcov with some minor changes. Note that those changes didn't
> modify any essential parts of lcov. My vague memory told me that they
> just made us life easier to parse the coverage results. We may not have
> our modified version any more.
>
> You can check some of our old scripts from:
>
> https://github.com/csmith-project/csmith/tree/master/utah/scripts/coverage
>
> Note we don't maintain those scripts. It's very likely they wouldn't be
> used for your purpose.
>
> Thanks,
>
> - Yang
>
>
>
Hi I used the script build_gcov_compiler.pl which made most of the
work of building the compiler easier.
Now heading towards gen_cov.pl script , the script has started to run
the compiler's testsuite for gcc , but it's stuck at

Running /home/extended_csmith/coverage/sources/current-gcc/gcc/testsuite/gcc.c-torture/compile/compile.exp
...

I have been running this script for more than half an hour, any
guesses how long does it take to finish the script? Or am I running
the script wrongly?

I have downloaded the lcov source code (a replacement to the hacked
lcov ) in the COV_BASE_HOME directory.

Thanks,
Sameeran Joshi