[2111] in SIPB bug reports
[daemon@ATHENA.MIT.EDU : Re: $HOSTTYPE in tcsh6 on the decmips]
daemon@ATHENA.MIT.EDU (ckclark@ATHENA.MIT.EDU)
Fri Sep 20 14:40:56 1991
From: ckclark@ATHENA.MIT.EDU
Date: Fri, 20 Sep 91 14:39:57 -0400
To: bug-sipb@ATHENA.MIT.EDU
Cc: jefft@ATHENA.MIT.EDU, marc@ATHENA.MIT.EDU
Reply-To: ckclark@mit.edu
I would almost be willing to say that HOSTTYPE should be a customizable
variable; i.e., that it should be something that the builder would be
able to tweak in a Makefile or configuaration file, rather than editting
tc.vers.c. Consensus on what HOSTTYPE should be is less important than
consensus on what `fs sysname' should return under AFS, for example. I
should also mention that /afs/sipb/project/tcsh/beta/src/tc.vers.c
currently has HOSTTYPE for the RS/6000 set to "rs6000". On an Athena
RS/6000, `machtype' returns rsaix, and this is the value it is set to in
the non-beta sources. If you plan on changing "decstation" to "decmips"
locally, then "rs6000" should be changed to "rsaix" by the same
argument. (The value of "mac2" for MAC AUX will also turn out to be
wrong, but I'm not going to push that point, considering the current
state of Athena MAC AUX.)
Conflicts over HOSTTYPE will affect the portability of csh scripts and
.cshrc files, but since it turns out that most scripts which have to
test HOSTTYPE are system-specific and inherently less portable anyway,
this may not be an important issue. However, the following line in the
README file for tcsh-6.00 seems to indicate that there's some desire to
get one holy algrorithm for setting HOSTTYPE once and for all, although
maybe I'm reading more into this than is there:
4) Look at sh.c to make sure that you have the right #ifdefs to set
the environment variable "HOSTTYPE" correctly. If you need to make
changes, PLEASE SEND THEM BACK TO ME.
-Calvin
