[Aranym-user] GNU/GCC 3.3.6 SDK for miniPack
Francois LE COAT
lecoat at atari.org
Mon Mar 24 23:16:57 CET 2014
I suspect we already had that kind of talk on Atari-coldfire list ?
Then I could perfectly build Eureka 2.12 with GNU/GCC 4.x, but
it was behaving totally wrong, and I couldn't tell you why ...
That's why I'm speaking of an incompatibility at sources level.
Eero Tamminen writes :
> Francois LE COAT wrote:
>> GNU/GCC 3.x is not better or worse than GNU/GCC 4.x ... It's just
>> different way of writing computer's languages. I would say that
>> 3.x is better if you write plain C language, and 4.x for C++ extension.
>> That's a matter of generations of developers ... I started writing
>> Kernighan and Ritchie C language, than I moved to pure ANSI C. The
>> transition to C++ for me is not natural, at all. Many warnings in
>> GNU/GCC 3.x became errors in 4.x. I can't cope with this 4.x style.
> What kind of "warnings become errors"?
>> The problem, when I build Eureka 2.12 with GNU/GCC 4.x, is that
>> it behaves completely wrong. Eureka 2.12 being a large piece of
>> C language, I can't tell you why, and I have no time to spend on it.
> What optimization options you're using?
> The higher optimization options you use, the more correct
> your code needs to be. Lower optimization levels are more
> forgiving of programming errors, and compiler's stricter
> interpretation of C specifications.
> For example, you might want to try build with "-fno-strict-aliasing".
> (Aliasing optimizations are enabled in GCC by -O2.)
>> So, I prefer building Eureka 2.12 with GNU/GCC 3.3.6. Perhaps it
>> could be useful for other developers, who would like to build old code.
>> It's a pity we can't access to GNU/GCC 3.3.6 cross-compiler, because
>> I would really enjoy using it, better than native, that is not so bad.
François LE COAT
Author of Eureka 2.12 (2D Graph Describer, 3D Modeller)
More information about the cz-bobek-lists-aranym-user