newlib vs C99 math functions

[mwoehlke]

Jeff Johnston wrote:
> Christopher Faylor wrote:
>> On Thu, Aug 10, 2006 at 03:49:01PM -0500, mwoehlke wrote:
>>> Christopher Faylor wrote:
>>>> On Thu, Aug 10, 2006 at 02:35:44PM -0500, mwoehlke wrote:
>>>>> Andrew Pinski wrote:
>>>>>> While porting some code over to use newlib from a closed source libc,
>>>>>> I ran into the problem of that newlib did not implement most of the
>>>>>> C99 math functions or macros.  I ran into the problem that FP_ILOGB0
>>>>>> was not defined and I was wondering if anyone is working on adding 
>>>>>> any
>>>>>> of the C99 math functions to newlib?
>>>>>>
>>>>>> Also is there a bug system for newlib because it would be great to
>>>>>> file bugs about the problems I ran into.
>>>>>>
>>>>>> Thanks,
>>>>>> Andrew Pinski
>>>>>
>>>>> Well, newlib's printf() does not support "%a"/"%A" which are 
>>>>> required by C99, so my observation of the C99 support at least is 
>>>>> "spotty at best".
>>>>>
>>>>> Meanwhile, I'll wait anxiously to see you get any more attention 
>>>>> from the 'official' folks than I've gotten so far.
>>>>
>>> I apologize for getting used to the promptness of Cygwin's list. This 
>>> one *does* seems dead by comparison.
>>>
>>> Now that we know you're here, do you know if anyone is working on 
>>> adding C99-conformance to newlib? ;-)
>>
>> I don't think that newlib has a roadmap.  Any changes would come from
>> either contributions or talking very nicely to Jeff.  Since newlib isn't
>> Jeff's full-time job, I would imagine that adding major new
>> functionality won't be high on his to-do list.
> 
> This seems as good a place to jump in as any now that I am back from my 
> vacation.

Ah, that explains why only Dave and CGF answered me... WB! Have a good 
time? :-)

> 2. There is no roadmap in place to be fully C99
>    or POSIX-compliant.  Newlib is still targetted at
>    embedded systems with a minimal set of syscalls.  That
>    said, I have no problem with adding C99 and POSIX features
>    to newlib that are helpful in getting more applications to run.
>    I recently added hex floating-point support to strtod for
>    example.

"Hex floating point support"... would that be the inverse of "%a"/"%A"? ;-)

In the mean time, would you be receptive if someone (me) took on 
overhauling the existing documentation system? (See the thread I FW'd 
here from cygwin-talk, "Re: What's wrong with *roff, anyway?".

I'll omit further comment as I seem to have commandeered this thread 
from its original purpose. :-) Further discussion to one of the threads 
I started, please!

-- 
Matthew
"We're all mad here. I'm mad. You're mad... You must be, or you wouldn't 
have come here." -- The Cheshire Cat