🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: gnu.emacs.bug
Article #976

compilation bug with emacs-15.58 on SGI 4D machines useing CC 3.1...

#976
From: Steve Parker
Date: Tue, 26 Jan 1993 02:16
54 lines
2832 bytes
To Whom it may concern,

I recently reported a problem compiling emacs-15.58 on an SGI personal
Indigo using their Ansi C compiler, version 3.10.

I also called SGI for help.  I got a reply but solved the problem myself
by getting emacs-15.59.tar.Z from prep.ai.mit.edu and installing that.
The problem was solved in this new version and the fellow at SGI gave this
to me as a possible explanation:

> The file `unexmips.c', in the src directory for the 18.58 version of
> emacs, will not work if the 3.10 SGI compilers are used to build this
> distribution.  This has to do with the order of the sections in the
> object file.  From what I can tell, this order is hardcoded in the
> 18.58 distribution.  This appears to be fixed in the 18.59 version (a
> loop is used instead to search for sections in an object file).  If you
> run into problems with temacs dumping the final executable under 18.58,
> and you are running the 3.10 compilers, please try to build 18.59
> instead where this problem is fixed.

> I believe that this explanation of why this error occurs is accurate, but I
> can't guarantee it.  I'm fairly positive, however, that installing 18.59
> instead will solve this problem.

> Krishna Sethuraman
> krishna@sgi.com

I thought that you could pass this information on to anyone using the new
Ansi C compiler from SGI.

And adjust the $EMACS/etc/MACHINES file to read:

---------------------------------------------------------------------------
Iris 4D (m-iris4d.h; s-irix3-3.h or s-irix4-0.h)

  18.59 is needed for SGI's Ansi C compiler version 3.10.

  18.58 is known to work on Silicon Graphics 4D series machines
  with IRIX 3.3 or IRIX 4.0.

  Most irix3.3 systems do not have an ANSI C compiler, but a few do.
  If you are using the ANSI C compiler, you may need to add
      #define C_SWITCH_MACHINE -cckr
  to config.h.

  There is a bug in IRIX that can sometimes leave ptys owned by
  root with a permission of 622.  This causes malfunctions in use
  of subprocesses of Emacs.  This may be fixed in IRIX 4.0.5.
---------------------------------------------------------------------------


 Steve Parker                    parker@mps.ohio-state.edu
 Dept of Chemistry               201 McPherson Lab.
 Ohio State University           614-292-5042
Technical Details
Message-ID: <9301252153.AA13652@sodium>
Path: arc1-ca.pugleaf.net!sparky!uunet!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!cis.ohio-state.edu!sodium.mps.ohio-state.edu!parker

Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!sparky!uunet!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!cis.ohio-state.edu!sodium.mps.ohio-state.edu!parker
From: parker@sodium.mps.ohio-state.edu (Steve Parker)
Newsgroups: gnu.emacs.bug
Subject: compilation bug with emacs-15.58 on SGI 4D machines useing CC 3.1...
Date: 25 Jan 1993 21:16:30 -0500
Organization: GNUs Not Usenet
Lines: 55
Sender: daemon@cis.ohio-state.edu
Approved: bug-gnu-emacs@prep.ai.mit.edu
Distribution: gnu
Message-ID: <9301252153.AA13652@sodium>