Re: QNX Port

["Andy Olson" <olsona at uswest dot net>]

Hmm, I know I can release it myself under the GPL, but I don't know how
"friendly" QNX will be to reassigning their copyright to what they
percieve to their entire toolchain.  Or they might be fine with it.

Now, it turns out that the main changes were these:

1) new config.{sub,guess} from GNU, w/ my very own patch

2) makefile fragments, config headers - these were COMPLETELY changed
from the QNX versions, and are now about 10-30 lines long each and look
more like the Solaris x86 ones.  Yes, I based the knowledge for how to
write them on QNX's work, but I started with a empty file, and added
either from "svr4.h" and other official headers, or that were my own
(turns out the QNX specs were singularly horked up).  The only things
that are similar to the QNX files are the -D flags, and that is
certainly a compatability issue.  Where these flags are set, and the
code that does it are not QNX's.

3) Code changes.  In the each of the "main" routines a call was added
that unlocks memory, allowing some swapping on this platform.  Also, in
one or two places in libstdc++ a couple lines had to be ifdef'ed out for
QNX (since the system C headers still managed to conflict).  Each of
these changes is so trivial that there is no other way to accomplish the
same (obvious) thing.

Do you think these qualify as changes to QNX's work, or would my own
copyright reassignment be sufficient.

Thanks,

Andy


"Joseph S. Myers" wrote:
> 
> On Fri, 18 Jan 2002, Craig Rodrigues wrote:
> 
> > If you could go to http://gcc.gnu.org/cvs.html, and try integrating
> > QNX RTP support into the latest GCC 3.1 tree, and submitting patches, that would be
> > really cool.  I know people who have had problems with GCC under QNX,
> > especially for C++.
> 
> Of course, if the patches are based on someone else's previous work on QNX
> support, you may need to get a copyright assignment from whoever did the
> previous work.
> 
> --
> Joseph S. Myers
> jsm28@cam.ac.uk