aros arm binutils generate real executables

[magorium]

Hi,


When I cross compile from *nix-64 to AROS-arm the binutils seem to create a real elf executable (instead of relocatable) and also reporting a different ABI version (1 instead of 0).


I'm unable to comprehend why this happens. Binutils patches and collect-aros originates from deadw00ds github. Did some change happened for hosted arm ?


Anyone out there that has any idea as of why ? and not less important how to handle this situation ?, because i'm stuck atm.


I would simply like to compile some things for AROS-arm.

[miker1264]

Hi,


When I cross compile from *nix-64 to AROS-arm the binutils seem to create a real elf executable (instead of relocatable) and also reporting a different ABI version (1 instead of 0).


I'm unable to comprehend why this happens. Binutils patches and collect-aros originates from deadw00ds github. Did some change happened for hosted arm ?


Anyone out there that has any idea as of why ? and not less important how to handle this situation ?, because i'm stuck atm.


I would simply like to compile some things for AROS-arm.

Just a guess. Michal Schultz has been working with AROS ARM development. The "1" is ABIv1 instead of ABIv0.

[magorium]

Just a guess. Michal Schultz has been working with AROS ARM development. The "1" is ABIv1 instead of ABIv0.

ah, thx. I see.


Ok, so the current code is a bit too hocus pocus for me atm but reverting to old svn aros tree and using the collect-aros from that, solved the issue for me now.


Note to self: current available collect-aros from github is unable to support standalone (cross) compilation for available hosted arm.

[deadwood]

Indeed Michal was working on changing the executable format in the main AROS repository. You mention using binutils patches from my repository - for clarity - arm is not maintained in my repository so I don't know how good or bad these patches are.

[magorium]

Thank you for the confirmation deadw00d.

I've not been keeping track for a longer period and right before that i had my scripts updated to use your github repo instead of using the old svn. So i was wrecking my brains there for a while 

I'll try to determine what differs and if it can easily restore the hosted support with a patch to the current code (if even necessary). The quickest for me now was simply to fall back to the sources from the old svn tree.