[236] in Kerberos-V5-bugs
Problems building R5 beta 2
daemon@ATHENA.MIT.EDU (Bob Beauchemin)
Sat Oct 3 14:03:11 1992
Date: Sat, 3 Oct 92 11:01 PDT
From: bobb@agora.rain.com (Bob Beauchemin)
To: krb5-bugs@Athena.MIT.EDU
I've been experiencing MANY problems trying to build Kerberos R5 beta 2.
Below is a step-by-step log of the problems I've been having (* indicates the
corrective actions I took). I finally had to make almost all the Makefiles "by
hand" and am concerned about possible problems I could introduce this way (I've
just started testing).
-------------------------------------------------------------------------------
*Restore all source to "/mit/krb5/src"
*Took imake and makedepend from Kerberos R4
*Compiled the latest com_err, ss, and unifdef stuff
*Compile isode release 8.0 (same as for Beta 1)
*Put imake, makedepend, unifdef, compile_et and mk_cmds in my path
*Changed site.def to point to my com_err, ss, and isode directories
*Because of the statements at the beginning of the main Makefile, I had to put
the distribution in /mit/krb5/src.
Made main Makefile, then...
making Makefiles in include
making all in /include/krb5
Imake.tmpl - no such file or directory
*Added setenv command,
setenv IMAKEINCLUDE -I/mit/krb5/src/config
Made main Makefile, then...
making Makefiles in include
making all in /include/krb5
Don't know how to make /mit/krb5/src/./stock/config.h
*Took out Srctop in site.def
Made main Makefile, then...
making Makefiles in include
making all in /include/krb5
...makes osconf.h and config.h
Don't know how to make Makefiles (in include subdir, because
include/Makefile -> include/Makefile.bak
*Made include/Makefile "by hand" by including correct stmts from main Makefile
*cd include
*make Makefiles
made /include Makefiles OK
*Took out subdir include in main Makefile and Imakefile
*at /mit/krb5/src, entered make Makefiles
making Makefiles in lib
...makes error_tables
making all in lib/asn.1 (starts to make asn.1 without generating any
Makefiles!)
*At this point, I MANUALLY made all the makefiles,
*Did "make depend" and "make all"
The "make world" procedure worked FINE with Kerberos R5 beta 1, which I
compiled last week (as a "practice run"). Am I missing something obvious, or
are other people having similar problems?
PS: I also had to copy the UNIV.py, UNIV.ph from the isode distribution so that
pepsy would correctly compile the "PasswdData" type in KRB5.py. I did not have
this problem in R5 Beta 1 (as there was no Passwd Data statement).
Thanks in advance for any assistance
Congratulations on the release of Beta 2!!!!!!!!!!
Bob Beauchemin
Open Computing Security Group
bobb@agora.rain.com
