What is "no" application? nm?

[Ralf Corsepius]

On Tue, 2008-04-01 at 09:04 +0200, Torsten Landschoff wrote:
> Ralf Corsepius wrote:
> >> Ooops... Just now I noticed that I use the so called 4.9 tool chain. Is this 
> >> incorrect and may be harmful?
> >>     
> > Difficult to answer.
> >
> > * rtems4.9 and the rtems4.8 toolchains' binutils are more or less
> > identical
> > * rtems4.9-gcc is at gcc-4.3.0, while rtems4.8-gcc is at gcc-4.2.x.
> > * rtems4.9-newlib is newlib-1.16.0, rtems4.8-newlib is newlib-1.15.0.
> >   
> IIRC gcc 4.3 is more aggressive in optimization wrt. memory barriers and
> such. I'd suggest using gcc 4.2 to be on the safe side.
Exactly. 

That's the reason why gcc-4.2.x is the "nominal toolchain" for "stable"
branch of RTEMS (aka. rtems-4.8) and why gcc-4.3 is the basis for the
"experimental/unstable" branch of RTEMS (aka. rtems-4.9).

On the other hand, so far, experiences with gcc-4.3.0 and rtems-4.9 are
quite promising. 

To go further, there are some architectures (in particular the m68k),
for which we are considering to upgrade the rtems-4.8 toolchain's gcc to
gcc-4.3.0, because known issues the current rtems-4.8 toolchains suffers
from appear to have been fixed.

> About the "no" application: With a probability of 90% this is the result
> of autoconf not finding a program which it thinks is not required. In
> that case, the program name may end up being the result from
Correct.

> checking for nm ... no
> 
> You might want to check your config.log and config.status files.
In 100% of all cases, I have seen so far, it has always been a screwed
up $PATH.

Ralf