[924] in Kerberos_V5_Development
Re: lib/crypto/Makefile.in changes
daemon@ATHENA.MIT.EDU (Theodore Ts'o)
Mon Sep 25 18:53:05 1995
Date: Mon, 25 Sep 1995 18:52:46 -0400
From: Theodore Ts'o <tytso@MIT.EDU>
To: Ezra Peisach <epeisach@MIT.EDU>
Cc: krbdev@MIT.EDU
In-Reply-To: Ezra Peisach's message of Sat, 23 Sep 1995 09:19:33 EDT,
<9509231319.AA29289@kangaroo.mit.edu>
Date: Sat, 23 Sep 1995 09:19:33 EDT
From: Ezra Peisach <epeisach@MIT.EDU>
a:) all: all-$(WHAT) has ^M's imbedded which causes gmake to have a fit
I couldn't find any more of such files this morning. I thought I had
gotten them all late friday night. I did forget to checkin some ^M
fixes in des_crc.c, des_md5.c, and raw_des.c, but I took care of that today.
b) Because you added that line, normal library building is broken for unix.
You will not descend into the subdirs before doing the all-unix...
The order int eh Makefile is
all: all-unix
all-unix: make the lib
all:: doi the subdirs
Perhaps we need to add the the subdir rule a different tag as well?
We had run into this problem before, the last time Fusion had checked
stuff in. The problem is that aclocal.m4 inserts rules that use all::,
but in the Makefile we were using all-unix::. This caused the ordering
problem. We had worked around it using some ugly hacks, which got
broken (for the PC and Mac) when we put in the shared library stuff.
Then Keith "fixed" things and broke it for Unix.
So I've finally fixed thigs right, which means that aclocal.m4 generates
rules that use only all-unix as a target. This applies for the subdir
recursion rules, as well as the shared library rules. The
"all: all-$(WHAT)" lines are now in config/pre.in, so that they are
defined for all makefiles.
Basically, the idea is to use all-unix:: everywhere, consistently. This
should hopefully keep the Makefile ordering headaches down to a mild
migraine....
- Ted
