Experimental Physics and
Industrial Control System

[email protected] (Low Master Kim) · Wed, 18 Jan 1995 11:06:47 -0700

All,
  I have worked on porting the first beta 3.12 EPICS to Solaris 2.4.  
My approach was to make the changes required to get the source to compile on 
the Solaris machine with the Solaris libraries and includes. It
is my view that using the ucblib and ucbinclude files as Andrew reports is 
_not_ a port of EPICS to Solaris - it is making do until a real port is
done.  I think it is important that the Solaris port not rely on the
compatibility libraries which may eventually go away.  I would urge the
system maintainers not to distribute a Solaris EPICS that needs the ucb
compatibility libraries.
  I worked through the source for base and several of the extensions including
dm and medm.  I have not sent my changes to Janet Anderson.  The list of
changes Andrew has is mine it is just notes I made to myself and doesn't
include diff files, etc.  If there is a mechanism available I would be
happy to work on a Solaris port that would be supported by the "home office."
  Very few substantial things need to be done to get EPICs to run on the
native Solaris system.  Most of the problems I ran into were caused by the use 
of BSD system calls that don't exist in the Solaris libraries (but do exist
in ucblib) such as bcopy and bzero, etc rather than memcpy, memset, etc.
and other BSD things like index, getwd. Solaris often requires different 
include files.  Also the SYSV install doesn't work like the BSD one (or
else it's a bug in Solaris) so I had to modify the .rules slightly.  I'm
not sure how the installation setup would handle different rules.  I got
the system to compile but the result is not heavily tested.
  If all developers used something standard like Posix the Unix-side code would
run on Sunos and Solaris out of the box.  The great thing is that there is
really very little that needs to change in the source (particularly the base)
to get to that place.  Some extensions were more difficult to get running.
It was a lot of time (a couple hours) to get dm to work.  Changing the source
and getting the base compiled under Solaris only takes a day or so.  Some
extensions I didn't bother with (knob manager) or I couldn't figure out the
Makefiles (not all the extensions used the new build scheme in R3.12 beta 1).
  I think it would be great if the consortium could discuss and agree upon
some common coding convention like Posix now that Posix is available and most
OS's support it.  There is a pretty good summary of Posix in an O'Reilly book
by Donald Lewine.  I think the changes to the core source would be minimal and 
the long gains to the project would be great.  Maybe this has been discussed
before.
  BTW, the new install procedures in R3.12 are simple and understandable - 
great job.
  Kim Gillies
  NOAO
  [email protected]
----- Begin Included Message -----

>From [email protected] Wed Jan 18 02:08 MST 1995
Date: Wed, 18 Jan 1995 09:08:10 GMT
From: Andrew Johnson <[email protected]>
Subject: Re: EPICS and Solaris 2.x
To: greene%[email protected]
Cc: [email protected]
Mime-Version: 1.0

On Tue, 17 Jan 95 13:40:03 MST greene%[email protected] wrote:

> Is anyone working on porting EPICS to Solaris 2.x?  If so, what's the
> time frame for that?

I have already 'converted' the R3.12 Beta to build and run under Solaris 2, 
using the Solaris ucb libraries (in fact there aren't too many changes needed 
once you get the CONFIG files right). Janet Anderson at APS is evaluating my 
changes and will be folding it into the official release. 

If you want a System-V version, I also have some of the changes required for the 
epics base, although these aren't heavily tested and need alterations to the 
CONFIG files or over-rides in the Unix makefiles.

- Andrew

          ,_}\_.        Andrew Johnson, Technology Division
         /      \       Royal Greenwich Observatory
~~~~~~~~| ,--/;. )~~~~  Madingley Road, Cambridge, UK
    .,  |/  (/  U       Phone: +44 (0)1223 374823
  .`., /_\ . .          Email: [email protected]
     wwWWWww'           WWW: http://cast0.ast.cam.ac.uk/~anj/

----- End Included Message -----

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System