[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [csmith-dev] csmith error

On 9/20/2011 10:12 AM, Eric Eide wrote:
> The checksum doesn't mean anything by itself.  But it is useful!
> The checksum is useful for finding compiler bugs through differential testing.
> In other words, you can compile your test program with several compilers --- or
> one compiler with several different combinations of options --- to see if the
> program always produces the same answer.
> Because of the way that Csmith produces tests, the checksum output of a test
> program should always be the same.  (Modulo a few details that I won't go into
> here.)
> For example, if you compile your test program with both "foobarcc -O1" and
> "foobarcc -O2", and the two compiled test program output different checksums,
> this is a very strong indicator that you have found a compiler bug.
> For a longer explanation of these ideas, I'll refer you to our published
> papers:
> 	http://www.cs.utah.edu/~regehr/papers/pldi11-preprint.pdf
> 	http://www.cs.utah.edu/~regehr/papers/emsoft08-preprint.pdf
> The go into more detail about how the generated programs work.
Thanks for the info.
> 	Bill> I still have troubles getting my compiler to compile the test
> 	Bill> codes. It crashed.
> If your compiler crashes, congratulations :-).  You've found a bug!  Maybe not
> the bug you were looking for, though :-(.
Thanks for your comforting words during this frustrating time. 
> 	Bill> I suspecting that typedefs are not correctly defined for my test
> 	Bill> setup. Can someone show me how to define types for my setup? The
> 	Bill> types for my system are: char 8 bits int 16 bits long 32 bits
> This seems fine; I don't see anything wrong with those representations of the
> standard types.
> Eric.
Thanks for clarification.