From: Sam S. <sd...@gn...> - 2002-09-27 18:26:54
|
> * In message <E17...@cm...> > * On the subject of "Re: Linux/m68k crash" > * Sent on Fri, 27 Sep 2002 18:15:34 +0100 > * Honorable Will Newton <wi...@mi...> writes: > > On Friday 27 Sep 2002 2:22 pm, Sam Steingold wrote: > > > type is the section that specifies the type of the object (fixnum vs > > char &c) > > data is the information in the object (the value of the fixnum &c) > > So what is the method for getting the data out? > AND with "mask" then shift by "shift"? see macros typecode, upointer, untype &c. > How many bits of type information are needed? 8. see "# Type bits:" in lispbibl.d (this is why CLISP fixnums are only 24 bits) > > I suggest that you stick with a tried a true configuration: there are 3 > > of them in lispbibl.d: `generic', `beos' and `LINUX_SPARC_OLD_GLIBC' > > (which also includes MC680X0/LINUX - your case, right?) > > one of them should work. > > generic does not work - segfault before startup as usual. > LINUX_SPARC_OLD_GLIBC is the current configuration. BeOS prints the > banner and then segfaults, so may be a promising way to go - there may > be more than one bug here. I'll keep trying and see if I can find a > working combination. thanks! > BTW as I understand it malloc on Linux is a wrapper around the brk() > syscall, so the behaviour of malloc and layout of malloced regions is > largely dependent on the kernel rather than the C library. I am not in a position to even talk about it. All I can say is that the same linux kernel (2.2) with glibc2.1 supported the LINUX_SPARC_OLD_GLIBC configuration, while it broke down with glibc2.2. -- Sam Steingold (http://www.podval.org/~sds) running RedHat7.3 GNU/Linux <http://www.camera.org> <http://www.iris.org.il> <http://www.memri.org/> <http://www.mideasttruth.com/> <http://www.palestine-central.com/links.html> Oh Lord, give me the source code of the Universe and a good debugger! |