jec@iuvax.cs.indiana.edu (07/14/87)
We just installed SR9.5.1 and our version of MAGIC (the VLSI
thing from Berkeley) seems to have broken. I appears to have broken
in the memory management section which leads us to believe that we
are observing the malloc/sbrk problem reported earlier. Has anyone
got a fix for this?
Also, our GNU emacs that had worked under SR9.3 now doesn't
work. It gets a segmentation fault after clearing the screen and putting
out a message about it not having anything loaded (which is true).
Thought I would add this to the list of known problem. Our Emacs is of
the mid 17.XX vintage.
III Usenet: iuvax!jec
UUU I UUU ARPANet: jec@iuvax.cs.indiana.edu
U I U Phone: (812) 335-5561
U I U U.S. Mail: Indiana University
U I U Dept. of Computer Science
UUUIUUU 021-C Lindley Hall
I Bloomington, IN. 47405
III (Home of the Indiana Hoosiers-- 1987 NCAA Basketball Champions)guzzi@uicsrd.csrd.uiuc.edu (07/15/87)
I can't help with MAGIC, but I do have a couple of comments
on gnu emacs. When we went from 9.2.5 to 9.5, my gnu emacs also
broke (18.26). I tried to recompile with the old C compiler
(/com/cc_sr9.2), but the libraries in /bsd4.2/usr/lib are of a
different format than that generated by the old C compiler and the
linker would not swallow it.
I then recompiled it under the 9.5 compiler without the
optimizer, and now it works (95% anyway).
I have heard that some very recent version of gnu has
been "fixed" to fit the apollo (version 18.44 or later?). You
might want to pick up the new version anyway since it has been
greatly improved since 17.xx.
--Mark Guzzi
University of Illinois at Urbana-Champaign
ARPA: guzzi@uicsrd.csrd.uiuc.edu
guzzi%uicsrd@a.cs.uiuc.edu
USENET: {siesmo,ihnp4,pur-ee}!uiucdcs!uicsrd!guzzidclemans@mntgfx.MENTOR.COM (Dave Clemans) (07/20/87)
The compilers released by Apollo with sr9.5.1 have some significant problems in the code optimizer. You may or may not hit them, but if you do all sorts of strange things can happen. The sr9.6 compilers are much better. About problems with malloc: remember that before 9.5 the Apollo memory allocation paradigm did not mesh well with what Unix-like programs expect (reference the "set sbrk size" kludge, among others). In 9.5 and later systems to match what a Unix-like program would expect; but I wouldn't be surprised if programs that intimately knew about the old system wouldn't run on 9.5 and later systems. dgc