Re: XCAR/XCDR as lvalues

Glynn> William M. Perry wrote: All bets are off with RedHat 4.x systems, unless you've built the libraries yourself. The version numbers are essentially meaningless; there was even some functionality in libc.so.5.0.9 that got omitted from RedHat's libc.so.5.3.12, IIRC, so it's not even monotone. Glynn> Right. Replacing Glynn> #include <sigcontext.h> with #include Glynn> <asm/sigcontext.h> Glynn> allows it to compile for libc-5 (for me, anyhow). Did it run? :-) What version of libc5? It didn't compile for my RedHat libc.so.5.3.12, but (a) it's a Sparc system, and that was RedHat's first shot at Sparc, and (b) my environment is possibly hosed (I'm transitioning to libc-2.1 by hand, I'm tired of poorly documented RedHat "enhancements", and to kernel 2.2.0, so there's chance my headers aren't getting picked up right, although I tried to make sure I got the right gcc, the right libc, and the right headers). But I don't think the libc5 issue is that important; if the signal-handling method gives efficiency improvements, we autoconf it when we see libc6 on Linux. People who continue maintaining libc5 systems clearly prefer stability to efficiency. This is a strong argument for being conservative about the test for purespace, though. -- University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091 __________________________________________________________________________ __________________________________________________________________________ What are those two straight lines for? "Free software rules."