[En-Nut-Discussion] Patch for SuSE 10.0 and gcc402

[Matthias Ringwald]

On 24.11.2005, at 12:46, Harald Kipp wrote:

>> Adding a "dont complain" compiler flag might fix the status quo,  
>> but the
>> real solution/improvement would be to do the tedious work and correct
>> all signedness differences.
>> I remember someone posting a patch to the sources a while ago
>
> You're right. But that patch had been quite voluminous.
> Looks like no one of the developers with CVS write access
> courageous enough to commit it.

Yes, now, I remember. This was exactly the reason I did hesitate
and did not just apply it. So I propose that we fix one signedness  
issue per week.. :)

Btw. I'm not using nut/os too much right now, but the next  
contribution from our side
probably will be another reduced critical section  
(NutThreadSwitch... :) and a new
radically optimized UART driver to mitigate our BT issue..
(only HW flowcontrol, only bocking read & writes, no (ring)buffer
and a IRQ single handler for both rx and tx to support max duplex  
speed ... )

>> harald, you might just add a line to the documentation (
>> "if using gcc > 4.0" add... ) and start fixing the sources..
>
> You're the Makefile script expert among us two. No better
> solution?

Hm.
a) I've completely lost track of the way nut/os is build by default  
(as we're using
my simple Makefile to support several targets in parallel).
b) if there's still a configure script that is called, one could add  
a simple check
to figure out the gcc version and add this option if gcc 4+

harald, can you give me a hint, where to add a gcc4 check? :)

cheers
matthias