Path: utzoo!mnetor!uunet!lll-winken!lll-tis!mordor!sri-spam!coco0.istc.sri.com!wohler
From: wohler@coco0.istc.sri.com (Bill Wohler)
Newsgroups: comp.sys.hp
Subject: problems with make
Message-ID: <12407@sri-spam.istc.sri.com>
Date: 4 May 88 01:36:03 GMT
Sender: nobody@sri-spam.istc.sri.com
Reply-To: wohler@coco0.istc.sri.com.UUCP (Bill Wohler)
Organization: SRI International
Lines: 45


  the hp9000s350 running 6.01 chokes on the following makefile
  fragment:

	$(LIBRARY):	$(OBJS)
			@echo -n "Loading $(LIBRARY) ... "
			@ar cru $(LIBRARY) $(OBJS)
			if [ "$(SYS)" != hp ]; then \
				ranlib $(LIBRARY); \
			fi
			@echo "done"

  my sun produces [newlines inserted for clarity]:

	% make SYS=hp
	Loading lib.a ... 
	if [ "hp" != hp ]; then \
		ranlib lib.a; \
	fi
	done

  but the hp spits out:

	% make SYS=hp
	Loading lib.a ...
	if [ "hp" != hp ]; then \
		ranlib lib.a; \
	fi
	Missing ].
	*** Error code 1

	Stop.

  but it's not the hp shell:

	$ if [ "hp" != hp ]; then ranlib lib.a; fi
	$ 

  note that i've had to create these "if" fragments to "script around"
  the way that hp deals with ar, ranlib and install.  the very polite
  ranlib message about ar already producing the SYMTAB should be
  suppressed and the install command is completely changed so that we
  need to use this thing called cpset.  when is using hp-ux going to
  become a pleasurable experience?

						--bw