135mb rpm packages and maybe 20mb tarballs waiting to be uploaded. Plus
a small pile of tarballs ... Problems still to solve:
- a change in the kernel makes ld.so die under certain circumstances.
- RedHat's sh-utils package is broken, it's pam support doesn't compile.
- When building the binutils 2.8.1-1 rpm cc1 dies with signal 6.
Strange, because the same package works great on the little endian
boxes and GCC actually shouldn't contain byteorder problems.
- the Emacs srpm refuses to he compiled on big endian machines
- I think I found the bug causing the large number of timouts the Seeq
driver is signalling. Haven't tested it, my kernel sources are in the
other end of the universe, under IRIX and I don't feel like rebooting ...
- timekeeping seems to be pretty broken, I saw my clock doing 10h in
three hours of uptime.
- the console code is somehow broken. Occasionally it writes garbage
characters under the lowest line that normally is used. When the
screenblanker is active one can see it writing garbage characters
on other places of the screen. Aside it's crawling, scrolling in less
reminds of a Sparc console of a slow Sparc. Furthermore I suspect
the console is also the cause of the memory corruption I occasionally
see.
Fixing #1 is the most important. It breaks recompiling libc as well as
the library dependency generation in RPM for all packages. RPM, btw. is
dumb and continues without noticing ...
Ralf
|