[7989] in Kerberos
Re: libcom_err.a on solaris? (B7)
daemon@ATHENA.MIT.EDU (Trever Furnish)
Mon Sep 16 22:15:09 1996
To: kerberos@MIT.EDU
Date: 16 Sep 1996 21:25:26 GMT
From: tfurnish@iupui.edu (Trever Furnish)
In an earlier article, I wrote the following, but I've found the problem...
In article <51k4q5$l1i@aim.et.iupui.edu> tfurnish@iupui.edu (Trever Furnish) writes:
>I'm trying to install beta 7 on solaris 2.5.1, but I keep getting the
>following errors on the make:
>
>false libcom_err.a error_message.o et_name.o init_et.o com_err.o
>make[2]: *** [libcom_err.a] Error 255
>make[2]: Leaving directory `/opt/local/krb5-beta7/src/util/et'
>make[1]: *** [all-unix] Error 1
>make[1]: Leaving directory `/opt/local/krb5-beta7/src/util'
>make: *** [all-unix] Error 1
>
>I thought error 1 was file not found, so I thought I'd try limiting make
>to one process next, but I've got absolutely no idea where to start for
>the libcom_err.a error. Neither have I been able to find a table of error
>numbers for Gnu gmake or gcc, which are what I'm attempting to build with.
The problem had nothing to do with kerberos, I guess... My path didn't
include the archive program, ar, and that was keeping libcom_err.a from
being generated. On solaris 2.5.1, that meant that I had to add /usr/ccs/bin
to the path of the user running the make.
Strange, you'd think things like yack, lex, ld, and make would've been
considered important enough by the folks at Sun to include in the default root
path.
Once the path was updated, the build worked.
However, one of the 'make check' tests failed -- the system doesn't return a
fully qualified domain name. (ie: I get bob, instead of bob.ind.net.) Could
someone post or point me towards the info on how to solve this problem? I
know it's been asked before, but unfortunately I didn't save the response.
Again, the platform is solaris 2.5.1 on an ultra spark, and again, any
help is greatly appreciated.
---
Trever Furnish,
trever@ind.net,
tfurnish@indyunix.iupui.edu
INDnet NOC
