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

Passing vector number (was Re: API modification request)

Sergei Organov wrote:
> Eric Valette <eric.valette at free.fr> writes:
>>Sergei Organov wrote:
>>>Please keep in mind that unfortunately you don't see all the code. There
>>>could be proprietary code that relies on features you change...
>>Using such argument, you will never change anything.
> No. It only implies that one should better think "I better don't change this
> unless I have valuable reasons to" instead of "Why don't change this if
> nobody seems to use it anyway?".

	1) there was a smiley in my message!
	2) I think we have come to the conclusion that a void* is indeed a more 
powerfull API with little performance impact (as you can cast the void* 
to an int to get your vector number). Concerning portability, as I 
already mentionned often in this newgroup, *IRQ HANDLING SHOULD REMAIN 
BSP SPECIFIC* meaning that you are not forced to change the API at all 
on non standart platform. If you just want to benefit the enhancement 
made by others, then this will cost you more and more time due to 
API/build enevironment cahnges that will occur.

> 1. There is very small reason to submit a code if it will only run on a
>    proprietary hardware that nobody can get anyway.

OK but if an API prevent to port rtems easilly to a hardware that is not 
too much exotic, maybe it means that the API is flowed...

> 2. Company policy may prevent submission. Most probably poor programmer bitten
>    by you hadn't established the policy :-(

Do you know Canon IP policy? Do you know the patenting pressure I had to 
fight all those past 7 years. So believe me, you can change things if 
you explain that submitting code cost less money and do not divulgate 
company know how...

   /  `                   	Eric Valette
  /--   __  o _.          	6 rue Paul Le Flem
(___, / (_(_(__         	35740 Pace

Tel: +33 (0)2 99 85 26 76	Fax: +33 (0)2 99 85 26 76
E-mail: eric.valette at free.fr