ken@gvax.cs.cornell.edu (Ken Birman) (01/08/90)
i put up isis 1.3.1 around here a couple months ago on our sparcstations running 4.0.3c and our nexts running next os 1?. i have a few non-technical suggestions that i've just gotten around to make: the location of log files should be compile-time and execution time configurable. where the isis processes run should be irrelevant. binaries on our systems tend to be run from read only nfs mounted filesystems. my rc files should not have to (cd /usr/spool/isis ...) log files seem to not have a unified name scheme and seem to be created with rather strange protections. on sparcs and nexts something similar to the following is found: total 22 -rw-rw-r-- 1 root wheel 20713 Jan 6 23:31 1.log drwxrwxrwx 2 root wheel 1024 Jan 6 21:34 logs/ -rw-r--r-- 1 root wheel 0 Jan 6 21:33 rexec.log --w--wx-wt 1 root wheel 0 Jan 6 21:34 xmgr.log* /usr/spool/isis/logs: total 1 -rw-rw-rw- 1 root wheel 212 Jan 6 21:34 XMGR-service on our nexts, even though we've fixed several of their brain damaged include files, our users have found the need to use -DNEXT -DMACH to compile, eliminating some inconsistancies between isis.h and stdio.h. the 1.0 release compiler seems to define -DNeXT and -D__MACH__. this may be documented in your manual, but i've not looked. and finally, will deceit be "freely available" just like isis is? -jay [ Ken: These seem like good suggestions; we'll look at including this ] [ Ken: sort of change into V2.0. This is a good time to raise issues ] [ Ken: like these, as we were just going to look hard at them. The ] [ Ken: xmgr mode looks to me like a bug, by the way. ] [ Ken: ] [ Ken: The need to specify -DMACH is unfortunate, this is because the ] [ Ken: isis.h include file didn't notice that it was on MACH and the ] [ Ken: declarations for system procedures differs from MACH to normal ] [ Ken: UNIX systems. Is there a predefined way to tell? If so, we ] [ Ken: can change isis.h to check. ] [ Ken: ] [ Ken: Our current plan is to make Deceit available when it runs well ] [ Ken: enough to be supportable by our group. Probably late in 1990. ]