still problem with new/old interrupts on PowerPC and CVS ?

Chris Johns chrisj at rtems.org
Sat Aug 6 01:58:43 UTC 2005


Peter Default wrote:
> 
> This relates to the issue of how RTEMS should appear to a given code  base.
> 

Yes.

> In my case, I'm trying to adhere to POSIX standards for real-time  
> threaded systems. I'd like to use RTEMS, and assuming no magic pre- 
> defines before including <unistd.h>, I'd like RTEMS to show up as a  
> POSIX threaded environment.

Agreed. RTEMS should follow the standards it supports.

> OpenBSD needs to show up, without something like BSD_ENVIRONMENT  
> defined before <unistd.h>, as a more BSD environment than I think  RTEMS 
> has to.

Sure. I do not think RTEMS should force people to follow a BSD 
environment. As Joel states examples or applications that use memory.h 
should be fixed.

I do not mind what happens with this file. The BSD header files have 
been a source of extra complication and anything that makes this cleaner 
or simpler is welcome. My point is we need to either have the file in 
RTEMS or document that this file has been removed and will never be 
accepted back into RTEMS.

A check of 3600+ FreeBSD source files I have handy shows the include 
only present in the source for a dos boot tool. Given this I see not 
reason to keep the file.

-- 
  Chris Johns



More information about the users mailing list