[Openembedded-users] gcc-4.2.4

[Pawel Osmialowski]

  Just shot in the dark: is it possible that weird behaviour of 
gcc-4.2.4 is caused by lack of gcc-flags-for-build.patch for this 
version in front of presence of zecke-no-host-includes.patch? I guess 
both deal with fixincludes feature of gcc-4.x which I hope that solves 
problem. I've just started to dig in gcc code structure, it is huge, so 
any help will be appreciated.

Cheers,
Paul


W dniu 12.09.2010 12:32, Pawel Osmialowski pisze:
>  Hello,
>
> I'm using recently obtained OpenEmbedded for building system image for 
> SH4 target. I desperately need to rebuild whole target system image 
> with gcc-4.2.4. So far it builds cleanly with gcc-4.4.4. 
> Unfortunately, I need to rebuild everything using 
> binutils-2.18.50.0.8+gcc-4.2.4+glibc-2.6.1. So far I successfully 
> downgraded binutils and everything is still building properly. 
> Unfortunately, when I downgrade gcc, I can't build glibc: both 
> current  2.10.1 and old 2.6.1 fails sanity check at configure stage: 
> it can't include limits.h. If my suspicion is right, it is caused by 
> zecke-no-host-includes.patch introduced recently. When I remove this 
> patch I'm running into worse troubles at later stages - I guess 
> currently OpenEmbedded relies heavily on changes introduced in 
> gcc-4.4.4 that helps to maintain barrier between build host includes 
> and build environment includes. Can you give me some clue what can I 
> do with gcc-4.2.4 to make it work with OE? If it is not possible, can 
> you give me some clue how can I get much older OE from the time of 
> gcc-4.2.4? I assume this gcc version was used with more or less 
> success by OE.
>
> Cheers,
> Paul
>
>
> _______________________________________________
> Openembedded-users mailing list
> Openembedded-users at linuxtogo.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-users