Xref: utzoo gnu.gcc.bug:1182 gnu.gcc:653
Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!agate!math.berkeley.edu!desouza
From: desouza@math.berkeley.edu (Paulo de Souza)
Newsgroups: gnu.gcc.bug,gnu.gcc
Subject: Bug on gcc-1.35 Makefile for the SUN3.
Message-ID: <25252@agate.BERKELEY.EDU>
Date: 5 Jun 89 00:41:45 GMT
Sender: usenet@agate.BERKELEY.EDU
Reply-To: desouza@math.berkeley.edu (Paulo de Souza)
Organization: Math Dept., UC Berkeley
Lines: 36

There is definitly a bug on gcc-1.35 Makefile for the SUN3.

I am trying to install gcc-1.35 on my SUN 3/50 and I am having  trouble
with one line of the makefile. After I do the configuring          with
/gcc-1.35/config.gcc sun3 all the links are set up properly and    then
when the makefile executes:
               "cc -c -g -I./config symout.c"
I get the following error message back:
              "symout.c: 1263: missing endif"
and it fails to build symout.o.
     
The line generating the error is line number 114 of the file      gcc-1.35/
config/tm-sun3.h (which is called in a sequence of includes) and is exactly
                #define DBX_DEBUGGING_INFO

If we comment out this line the compilation process finishes without     a
hitch including the second compilation, etc... . But the compiler     then
generated does not have the flag -g and it does not pass that   comparison
test between the first and second generation compilers that are  suggested
on the instalation file.

I do not fully understand what are the consequences of comenting out  that 
line, but from what we seen above it is not a really perfect compiler.  On
the other hand I saw that for some other machines this line of the corres-
ponding file is commented out. I need the gcc compiler in order to   build 
a Lisp compiler (Kyoto CL) and them compile a big and complex lisp   work,
so I would like to have the cleanest possible compiler. 

I would like to hear from people with the same problem and what fixes they
could come up with. I further state that the copy from all 3 machines that
I looked at (prep.mit, athena.mit, uunet)  has the problem (the copies are
exactly the same). I am running SUN OS 3.4 on a machine with floating poin-
ting hardare. If you could be of any help please come forward.


Paul de Souza                                   desouza@math.berkeley.edu