path for patches

[Ralf Corsepius]

On 04/11/2013 09:56 AM, Sebastian Huber wrote:
> I would not use GCC 4.4.7 on ARM due to this bug:
>
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=38644
>
> At least you have to make sure to disable this optimization.

Well, I think you are overdramatising. All tools/programs will always 
contain bugs somewhere - Bug free SW is an illusion. It's just a matter 
of these bugs to show effect in your personal use case.

That said, yes, it might be true, this bug might have affected your 
work, and it might be true, this bug may have disqualified gcc-4.4.7 for 
your work.

However, it this bug was as dramatic as you seem to be trying to make it 
appear, it would likely have seen more attention from the RTEMS users - 
This apparently wasn't the case.

Anyway, the real problem behind all this, is not this particular bug, is 
gcc-4.4.x being unmaintained and dead upstream, i.e. this bug won't ever 
be fixed for gcc-4.4.x, and rtems-4.10 not being worked at either (== 
dead, IMO)

> In general for ARM I would use the current RTEMS development version
> (future 4.11).
This is what I would have recommended some years ago, but ATM, I do not 
second this recommendation. IMO, rtems 4.11 is suitable as basis for OS 
development, but is way too volatile to be suitable as basis of 
application development.

Similar things can be said about gcc-4.8.0. It's too young to have a 
general opinion on it or of its usability for particular target.
The only difference to its predecessors is it, some of the known bugs 
its predecessors had suffered from, are _supposed_ to have been fixed.
This does not mean it carries less/more bugs or new/yet unknown bugs.

Ralf