[Mesa-dev] [PATCH 2/7] mesa: move FLT_MAX_EXP to c99_math.h

Brian Paul brianp at vmware.com
Sat Feb 28 07:36:58 PST 2015


On 02/28/2015 03:48 AM, Emil Velikov wrote:
> On 28/02/15 00:47, Brian Paul wrote:
>> ---
>>   include/c99_math.h       | 4 ++++
>>   src/mesa/main/compiler.h | 4 ----
>>   2 files changed, 4 insertions(+), 4 deletions(-)
>>
>> diff --git a/include/c99_math.h b/include/c99_math.h
>> index 35173c6..0a49950 100644
>> --- a/include/c99_math.h
>> +++ b/include/c99_math.h
>> @@ -156,5 +156,9 @@ llrintf(float f)
>>   #define M_LOG2E (1.4426950408889634074)
>>   #endif
>>
>> +#ifndef FLT_MAX_EXP
>> +#define FLT_MAX_EXP 128
>> +#endif
>> +
>>
> Strictly speaking one can move it to c99_float.h, but I may be nitpicking too much.

Yeah, I didn't think it was worthwhile to create c99_float.h just for that.


> Upon closer look at msdn [1] it seems that MSVC 2005 and later have the
> define. For the rest of the compiler world, I would assume that they're
> all C99 capable/compatible.

I'm not that confident.  Between MinGW and Vinson's *BSD testing, I 
wouldn't be surprised if it's missing somewhere.  I'd rather just play 
it safe for now.

BTW, my overall goal here isn't to update all the macros/includes/etc 
for modern compilers.  My priority is decluttering and refactoring to 
make things more organized.  Other clean-ups can follow.

Thanks for reviewing.

-Brian



More information about the mesa-dev mailing list