I suspect there is something not quite right in 1.3.37. I have seen two
total system hangs under it, under radically different hardware. In both
cases X was running, but that's the extent of the similarity. I used
1.3.35 enough that I think the problem was most likely introduced in
1.3.36 or 1.3.37; but take that with a grain of salt.
Machine A Machine B
Intel 486 DX2/66 AMD DX4/100
EISA PCI/ISA/VLB
1 IDE drive 2 IDE, 2 SCSI (NCR810 PCI)
NE2000 clone (used heavily) NE2000 clone (rarely used); PPP
Hung after about 5 days Hung after about 5 hours
Running xlock -mode mountain Using several xterms interactively
Compaq Qvision 1024/E 1MB Diamond S3 Trio64 2MB
Xinside AcceleratedX XFree86 3.1.1/2 mixture
Mostly ELF, some a.out Mostly ELF, some a.out
GCC 2.7.0 GCC 2.7.0
libc 5.2.10 libc 5.2.10
While some people would be inclined to point out the NE2000 clones, I must
caution them that it is exceedingly rare for Machine A to hang or crash
under any circumstances. Since 1.2.6 it typically stays up for weeks at a
time, and it has only been rebooted due to power failures, hardware
changes, kernel upgrades, and (once or twice) bad NFS situations. It is
of course possible that these two hangs are coincidence, or that the
problem is merely *exacerbated* by 1.3.37 as opposed to *caused by* it.
---David C. Niemi---niemidc@clark.net----703-904-3596---Reston, VA USA---
There is no match for the fervor of a new initiate, the confidence of the
unquestioning, or the appeal of oversimplified answers to complex issues.