[Stackless] stackless clang woes....

Jeff Senn senn at maya.com
Fri Sep 20 18:48:22 CEST 2013


That was not it -- in fact the ARM port was still fine (though I do use a slightly different toolchain for that) -- just the 32 bit version of i386 for OS-X was misbehaving.  I suspect it had to do with some inlining or optimization that clang was doing (differently than a previous version).

I applied some of the recent changes (including an optimizer-defeating, write-only variable hack in slp_transfer.c which it appears that Anselm added) and the problem seems to have resolved. Thanks, Anselm!

On Sep 19, 2013, at 4:20 PM, Richard Tew <richard.m.tew at gmail.com> wrote:

> I remember when I was compiling Stackless for the Nintendo DS, there
> was a version of GCC which didn't accept the standard inline assembler
> to save and restore the stack pointer.  I had to write a full
> assembler method for slp_switch.
> 
> See: switch_arm_thumb_gas.s
> 
> Cheers,
> Richard.
> 
> _______________________________________________
> Stackless mailing list
> Stackless at stackless.com
> http://www.stackless.com/mailman/listinfo/stackless
> 




More information about the Stackless mailing list