On Sun, Jul 11, 2010 at 11:10:29PM +0100, Max Bolingbroke wrote:

On 11 July 2010 22:35, John Meacham wrote:

...

Hi, I know there is an issue on some vesions of OSX due to the lack of posix_memalign, but this appears to be a different thing.

Yep, I'm on 10.6 which apparently has memalign.

...

Can you send me the output of 'uname -a' and 'gcc --version' and what version of OSX you are using? thanks.

Same as in my reply on -cafe: OS X 10.6.4 and:

$ uname -a && gcc --version Darwin REDACTED.ac.uk 10.4.0 Darwin Kernel Version 10.4.0: Fri Apr 23 18:28:53 PDT 2010; root:xnu-1504.7.4~1/RELEASE_I386 i386 i686-apple-darwin10-gcc-4.2.1 (GCC) 4.2.1 (Apple Inc. build 5659) Copyright (C) 2007 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

Hi, out of curiousity, could you try editing the .c file and adding #define _start start #define _end end after the header file includes and before the extern void _start,_end; declarations then try recompiling it? then if that doesn't work, try #define _start __start #define _end __end The command you need to run to recompile the .c file will be embedded in the file near the top of it, or the utils/recomp.prl script will pull it out and recompile a .c generated by jhc. (useful when debugging the rts). I know that OSX uses different rules for linking and whether underscores are prepended, but I forget the details. John -- John Meacham - ⑆repetae.net⑆john⑈ - http://notanumber.net/

On Sun, Jul 11, 2010 at 11:56:01PM +0100, Max Bolingbroke wrote:

Are you expecting the _start and _end labels to be defined in the .s file output by GCC? I can't see anything like them in the -S output. If not, what library do they originate from?

Neither, they are generated by the linker on linux, I guess they just don't exist on OSX.

Is there any other information you might be able to use to figure out this bug?

No, I was just hoping one of those might be an easy fix. The _start,_end mechanism is pretty much a hack anyway, I was going to replace it with a more reliable mechanism in any case. John -- John Meacham - ⑆repetae.net⑆john⑈ - http://notanumber.net/