[csmith-dev] disable 8-bit integer types

[Tuyen Hoang]

Hi Yang,

Thanks for your very quick response, a lot quicker than I expected. And I am happy to tell that the added options work for me. I really appreciate that.

Just one more favor, though. If I should need to hand-write some tests, I'd prefer to use same self-checking (checksum) mechanism employed by Csmith's generated tests. Could you please point me to some sources for more implementation details?

Regards,
Tuyen



>________________________________
> From: Yang Chen <chenyang at cs.utah.edu>
>To: Tuyen Hoang <tuyenqh at yahoo.com> 
>Cc: "csmith-dev at flux.utah.edu" <csmith-dev at flux.utah.edu> 
>Sent: Saturday, June 22, 2013 12:10 AM
>Subject: Re: [csmith-dev] disable 8-bit integer types
> 
>
>Hi Huyen,
>
>I added two options for your purpose:
>
>csmith --no-uint8 --no-int8
>
>You can checkout the latest Csmith from our git repo.
>
>Note that our runtime relies on uint8_t to compute crc32 checksums. You 
>may need to manually change that part in runtime/csmith.h
>
>But with the above options, Csmith should be able to generate programs 
>which do not contain int8_t and uint8_t types.
>
>- Yang
>
>Tuyen Hoang wrote:
>> Hello group,
>>
>> I'm considering using CSmith to test our DSP C compiler, and one 
>> problem is currently preventing us from doing using it: our compiler 
>> does not support 8 bit integer types (as well as 'long long'). There 
>> is already an option to disable generating of long long entities, 
>> would it be easy to add one for 'byte'?
>>
>> Because of time pressure, I'll try to modify the source code myself 
>> while waiting for your response, but I'm afraid that, although it is a 
>> simple thing for you to do,  it would take me considerably longer time.
>>
>> So please help me.
>>
>> Regards,
>> Tuyen
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: </listarchives/csmith-dev/attachments/20130624/adf81bd1/attachment-0002.html>