[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [MiNT] [Atari-coldfire] Eureka 2.12 for Firebee



the fact that (most of) these issues look like stack related, but are
not, makes me think on the recent (12 months ago) changes to the CPU
registers before/after function calls, to make them the same for
PC/AHCC and GCC, and the changed use of return register for FPU
functions

this would mean the libraries used are not up to date, with latest CVS
header files etc

If the above was only for mintlib and mintlib is not being used, then
I guess this is a null post

Paul

On Thu, Mar 15, 2012 at 1:12 PM, Francois LE COAT <lecoat@atari.org> wrote:
> Hi Vincent,
>
> Vincent Rivière writes :
>>>
>>> But the real trouble is using another compiler. Some old code working
>>>
>>> well with GCC 2.x has problems with GCC 4.x because it is more strict,
>>> they
>>> need to be fixed.
>>
>>
>> BTW, using AHCC for ColdFire is another option.
>
>
> That's sensible :-)
>
> Eureka 2.12 was successively built with Lattice C, Borland C, Pure C 1.1
> GNU/GCC 2.8.1, 2.95.2, 3.3.3 and 3.3.6. And now GNU/GCC 4.5.2 ... It can
> still be built with Pure C 1.1, that is a reference tool hopefully.
>
> In my previous post you should read GNU/GCC 4.5.2, not 2.5.2, my fault.
>
> Now, a good point is that floating point computations seem to work.
> I only have a short experience with AHCC for the moment. It is nice !
>
>
>> But you are very close to a perfect Eureka for ColdFire compiled with GCC.
>
>
> You're encouraging :-) Thanks a lot. As a matter of facts, neither
> Pure C with 2nd level of warnings, nor GNU/GCC with -Wall are
> complaining about serious warnings. My conclusion, because meanwhile
> I also increased the executable stack-size, is that some system calls
> are not behaving the same way as it used to. Maybe some calls (bios,
> gemdos, aes, vdi ...) changed the way it acts with the GEM interface.
>
> It is not very simple to be sure about what's right or wrong, because
> I use a cross-compiler, and test it with ARAnyM. It would be much
> simpler with a native GNU/GCC compiler, and a real Coldfire hardware.
>
> The behavior of Eureka 2.12, I just built for you is rather chaotic.
> I mean, it works nicely about 1 minute or 2, and becomes hazardous.
> It really made me think that it was a stack corruption, but it's not.
> I don't make usage of USERDEFS, in dialogs or menus. It is just a
> basic GEM interface, that works from early TOS up to recent systems.
>
> Don't be shocked, but it will take some time to fix it. Good challenge !
>
>
> Best regards,
>
> --
> François LE COAT
> Author of Eureka 2.12 (2D Graph Describer, 3D Modeller)
> http://eureka.atari.org/
> http://fon.gs/atarians/
>
>
> _______________________________________________
> Atari-coldfire mailing list
> Atari-coldfire@lists.lnxnt.org
> https://lists.lnxnt.org/mailman/listinfo/atari-coldfire