Re: 64-bit problem on Mac

Hi Alex, Hi Jon, I now re-did everything from scratch, in a new clean ongoing test, and I think the results are now more what you had expected. The error messages were now just this: sed: 1: x86-64.darwin.base.s: extra characters at the end of x command Ah, OK, this makes sense. It

Re: 64-bit problem on Mac

Hi Jon, Using sed -i'' 's/@plt//' x86-64.darwin.base.s did not delete the @plt Hmm, so let's give up inplace-editing with 'sed'. We can use three commands instead: Move the '*.s' file to some temporary file, and then 'sed' to remove the '@plt's: mv x86-64.darwin.base.s

Re: 64-bit problem on Mac

Hi Jon, I now re-did everything from scratch, in a new clean ongoing test, and I think the results are now more what you had expected. The error messages were now just this: sed: 1: x86-64.darwin.base.s: extra characters at the end of x command Ah, OK, this makes sense. It seems that 'sed'

Re: 64-bit problem on Mac

Hi Jon, and send me the resulting main.s? ... See attached file. Thanks! Looking at it, it seems a long way to go ... Besides the @plt issue, there are other differences. For example, symbol names are prefixed with an additional underscore. But let's start with the first step, removing the

Re: 64-bit problem on Mac

Hi Jon, Sorry for the delay, but here are the errors I got after I appended the 'sed' calls (see attachted file). I'm not shure if there's a big difference, but you may spot something ... Indeed. No difference at all ... x86-64.darwin.base.s:1957:junk `@plt' after expression

Re: 64-bit problem on Mac

Hi Jon, I've now got an answer to my question at http://stackoverflow.com/questions/6384961/osx-gnu-assembler-problem-with-call-fooplt Nice. Matthew Slattery says: .. 1. PLT is an ELF concept, but OS X uses a completely different object / executable file format - Mach-O.

Re: 64-bit problem on Mac

Hi Alex, Hi Jon, What helps in such cases is usually writing code snippets in C, compile them to assembly code (the -S option to gcc), and study that. .. I think this sounds like something to try. Do you have any suggestions for code snippets in C that can be useful? OK. Perhaps the most

64-bit problem on Mac

Hi, I wanted to look at the old problem of building 64-bit PicoLisp on Mac OS X. Assemblers are far from what I usually play with, but at least I'd like to spend a few hours today on this problem. From what I've found on the net, I got the impression that the standard (?) assembler on OSX is

Re: 64-bit problem on Mac

I thougt PicoLisp used its' own assembler? On Thu, Jun 16, 2011 at 11:24:05AM +0200, Jon Kleiser wrote: Hi, I wanted to look at the old problem of building 64-bit PicoLisp on Mac OS X. Assemblers are far from what I usually play with, but at least I'd like to spend a few hours today on

Re: 64-bit problem on Mac

Hi Alex, Hi Jon, I wanted to look at the old problem of building 64-bit PicoLisp on Mac OS Great! :) .. I hope this gives you some hints to get started. We can try to tackle this together once we find out what exactly goes wrong. Cheers, - Alex It's probably best to continue this on

Re: 64-bit problem on Mac

Hi Jon, It's probably best to continue this on IRC. I have some other business for about one hour. I'll log in when I'm ready. OK, fine. See you! Cheers, - Alex -- UNSUBSCRIBE: mailto:picolisp@software-lab.de?subject=Unsubscribe

Re: 64-bit problem on Mac

Hi Jakob, I thougt PicoLisp used its' own assembler? Well, yes and no. The assembly language PicoLisp is written in, is its' own. It is described in http://software-lab.de/doc64/asm; (language and virtual CPU architecture). The build procedure (triggered by the script src64/mkAsm) translates

Re: 64-bit problem on Mac

On Thu, Jun 16, 2011 at 12:02:23PM +0200, Alexander Burger wrote: As far as I remember, the only remaining problem was not that of the assembler, but of the linker, or, more correctly, the final output file format. The last time we discussed it here was: