Some further notes about util-vserver's failure to build from source: * dietlibc-dev on hppa fails to define O_FOLLOW and O_LARGEFILE. Adding these to CFLAGS allows util-vserver to build until it fails to link, because the linking looks for /usr/lib/diet/lib-hppa instead of /usr/lib/diet/lib-parisc, which dietlibc-dev provides. I believe, but have not tested (since it would require root) that if that is fixed in dietlibc, util-vserver will build on hppa.
* Similarly to hppa, everything should work on mipsel after dietlibc-dev provides /usr/lib/diet/lib-mipsel instead of /usr/lib/diet/lib-mips (or after libtool is changed to use the .../lib-mips directory). * Steve Langasek reported on IRC that adding -mv8 to CFLAGS should make util-vserver build on sparc. This means util-vserver won't work on the oldest sparcs, but we (or the kernel and glibc) don't seem to support them anyway. * m68k shouldn't build util-vserver at all, since dietlib-dev does not exist on that platform. I have mailed the Package-arch-specific maintainers that util-vserver should be added to that file to prevent build daemons from building util-vserver in vain (forgot to Cc Ola, sorry). * I couldn't replicate the build problem on trex.debian.org, which is an s390 machine. I assume the problem was transient. I can provide a patch for debian/rules for conditionally adding arch-specific stuff to CFLAGS, but later tonight or tomorrow. Again, I hope this is helpful. Happy hacking. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
