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

Re: [csmith-dev] Fwd: csmith errors



Hi Bill,

Just be aware timeout under Cygwin is not reliable - sometimes it doesn't kill Csmith after specified timeout limit, and the testing process is stalled. That's my experience with timeout under Cygwin. If it works for you, that's great. I'd like to hear about it because I have wanted to test on Windows for a long time.

-Xuejun
 
On 9/23/2011 4:05 PM, John Regehr wrote:
> Hi Bill, it is expected that some Csmith outputs hang.  You need to 
> run them under a timeout.  In Linux, this is easy to accomplish using 
> ulimit or similar.  Not sure about Windows.
>
> John
>
>
Hi John,
Thanks for the info. I ran them with timeout in my script in cygwin. I also want to make sure that you know that I'm not complaining, O:-)

Thanks and have a nice weekend.

Bill

>
> On 9/23/11 2:02 PM, Bill Chan wrote:
>> Hi John,
>> Thanks.
>>
>> But I have other issues here with some test codes generated by csmith.
>> When I run the executable, it hangs.
>>
>> Here is the gcc output of one of the problem test codes:
>>
>> gcc a2p_test5_msd_5_med_5.c
>> a2p_test5_msd_5_med_5.c: In function ‘func_36’:
>> a2p_test5_msd_5_med_5.c:484:17: warning: large integer implicitly 
>> truncated to unsigned type
>> a2p_test5_msd_5_med_5.c:514:38: warning: comparison of distinct 
>> pointer types lacks a cast executing a.exe...
>>
>> As you can see that a.exe hung. I attached the 
>> a2p_test5_msd_5_med_5.c here, see if anyone can duplicate the problem 
>> or spot what's wrong with the code.
>>
>> Some test codes with the same warnings have no problems finishing the 
>> execution, always output checksum.
>>
>> Thanks,
>>
>> Bill
>>
>>
>> On 9/23/2011 3:39 PM, John Regehr wrote:
>>> Hi Bill,
>>>
>>> There is only one question to ask about a program generated by Csmith:
>>> Does the C standard assign an unambiguous meaning to it?  If yes, 
>>> then testing can proceed.  If no, then we have to fix a bug.
>>>
>>> The warnings you list are designed to help programmers avoid 
>>> error-prone constructs.  Csmith is not a programmer, and the 
>>> warnings are irrelevant.
>>>
>>> John
>>>
>>>
>>>
>>> On 9/23/11 1:35 PM, Bill Chan wrote:
>>>> Hi all,
>>>>
>>>> Can someone explains to me what these gcc warnings mean?
>>>>
>>>> comparison of distinct pointer types lacks a cast large integer 
>>>> implicitly truncated to unsigned type overflow in implicit constant 
>>>> conversion
>>>>
>>>> I keep getting these warning when I use gcc to compile the test codes.
>>>>
>>>> Thanks,
>>>>
>>>> Bill
>>>>
>>>>
>>>> -------- Original Message --------
>>>> Subject:     csmith errors
>>>> Date:     Thu, 22 Sep 2011 15:43:34 -0400
>>>> From:     Bill Chan<billchan01@gmail.com>
>>>> To:     csmith_list<csmith-dev@flux.utah.edu>
>>>>
>>>>
>>>>
>>>> Hi all,
>>>>
>>>> I created testsuites with different levels of max_expr_depth. I 
>>>> compiled these files with gcc, and got the following warnings.
>>>>
>>>> comparison of distinct pointer types lacks a cast large integer 
>>>> implicitly truncated to unsigned type overflow in implicit constant 
>>>> conversion
>>>>
>>>> Attached is one of the 7 sets of test codes created. If anyone is 
>>>> interested in all of them, i can post them all later.
>>>>
>>>> -Bill
>>>>
>>>>
>>>
>