From: Timothy M. <mo...@la...> - 2003-11-24 12:02:11
|
Hello, I'm having some trouble compiling the OpenGL bindings from http://www.ii.uib.no/~knute/lisp/lisp.html with today's sbcl. Here's what I did: built sbcl with :sb-ldb :sb-thread loaded in clocc/clocc and clocc/src/defsystem-3.x/defsystem from a somewhat recent clocc compiled and loaded cmucl-xlib-and-gl.system (make:compile-system :gl) When running this under gdb I see: fatal error in thread 0x0, pid=25742 Hope this helps, Tim |
From: Christophe R. <cs...@ca...> - 2003-12-10 10:49:52
|
Timothy MOORE <mo...@la...> writes: > Hello, > I'm having some trouble compiling the OpenGL bindings from > http://www.ii.uib.no/~knute/lisp/lisp.html with today's sbcl. Here's > what I did: > > built sbcl with :sb-ldb :sb-thread > loaded in clocc/clocc and clocc/src/defsystem-3.x/defsystem from a > somewhat recent clocc > compiled and loaded cmucl-xlib-and-gl.system > (make:compile-system :gl) > > When running this under gdb I see: > fatal error in thread 0x0, pid=25742 I wasn't running under gdb, but using a lightly-modified sbcl-0.8.6.32 with :sb-thread, (make:compile-system :gl) has just run to completion. I did have to edit cmucl-xlib-and-gl.system to set the path to the source (how quaint :-), but apart from this (and that I don't have :sb-ldb, but I really can't believe that matters) it compiled fine. (I'm not about to test execution, I'm afraid; not from Windows over two ssh links :-) I don't know what to suggest if you're still seeing this. If you are, probably a complete session log would be most helpful. Cheers, Christophe -- http://www-jcsu.jesus.cam.ac.uk/~csr21/ +44 1223 510 299/+44 7729 383 757 (set-pprint-dispatch 'number (lambda (s o) (declare (special b)) (format s b))) (defvar b "~&Just another Lisp hacker~%") (pprint #36rJesusCollegeCambridge) |
From: Daniel B. <da...@te...> - 2003-12-10 16:08:22
|
Timothy MOORE <mo...@la...> writes: > Hello, > I'm having some trouble compiling the OpenGL bindings from > http://www.ii.uib.no/~knute/lisp/lisp.html with today's sbcl. Here's I think I caught the tail end of IRC discussion about this, and would like to restate for the record: with sb-thread we use %fs for thread-local storage. It's my belief that some graphics drivers (NVidia notably) also want %fs for their own purposes, and it's not completely impossible that building opengl stuff is where we'd see problems. Tim said his card is an ATI Rage 128. I can't find anything out about what drivers are needed for this nor whether they need %fs, but if anyone can shed more light on this aspect and either confirm or exclude it as a problem cause, that'd be pretty useful. -dan -- http://web.metacircles.com/ - Open Source software development and support |
From: Timothy M. <Tim...@la...> - 2003-12-10 22:15:42
|
Hi Dan, On Dec 10, 2003, at 5:07 PM, Daniel Barlow wrote: > Timothy MOORE <mo...@la...> writes: > >> Hello, >> I'm having some trouble compiling the OpenGL bindings from >> http://www.ii.uib.no/~knute/lisp/lisp.html with today's sbcl. Here's > > I think I caught the tail end of IRC discussion about this, and would > like to restate for the record: with sb-thread we use %fs for > thread-local storage. It's my belief that some graphics drivers > (NVidia notably) also want %fs for their own purposes, and it's not > completely impossible that building opengl stuff is where we'd see > problems. > But I'm not actually calling any OpenGL functions yet, just compiling and loading the C interfaces and then compiling the Lisp code. I think it's more likely that something is wrong with the way the .o files are compiled and loaded. I'll investigate tomorrow. Tim |
From: Daniel B. <da...@te...> - 2003-12-11 03:34:02
|
Timothy MOORE <Tim...@la...> writes: > [ %fs ] > But I'm not actually calling any OpenGL functions yet, just compiling > and loading the C interfaces and then compiling the Lisp code. I think > it's more likely that something is wrong with the way the .o files are > compiled and loaded. I'll investigate tomorrow. I wouldn't discount it as a possibility on that basis: with suitably inappropriate ELF magic one can specify bits of code in a library that get run on dlopen() - this is legitimately(sic) used for C++ constructors, for example. Still, I haven't yet found anything to suggest that the ATI drivers use %fs anyway. What linux distro, kernel version, libc version, etc etc? -dan -- http://web.metacircles.com/ - Open Source software development and support |