Experimental Physics and
Industrial Control System

"J. Lewis Muir via Tech-talk" <[email protected]> · Fri, 10 Jan 2020 14:34:06 -0600

On 01/10, Ralph Lange via Tech-talk wrote:
> On Fri, 10 Jan 2020 at 18:04, J. Lewis Muir via Tech-talk <
> [email protected]> wrote:
> 
> > [...] In practice, what this means is:
> >
> > 2. Creating and using binary packages is awkward at best because there
> >    is no way to upgrade a binary package safely because any new version
> >    (even a bump of just the least significant version number) can break
> >    ABI compatibility.
> >
> 
> In the context of the 22 releases of CODAC Core System, ITER has
> distributed EPICS as binary packages for 10 years to users in over 60
> organizations worldwide, where it was installed on several hundred
> machines, for development and production use.
> There was not a single situation (that I know of) being awkward for any of
> our users because EPICS Base does not promise ABI compatibility between
> versions.
> Are we living in the same universe?

Apparently not. :-)

I don't know how you're doing it, so I'd be very interested to hear if
you have time to briefly share.  But I would guess one or more of the
following makes it work for you:

1. Users only use the command line tools that come with EPICS.

2. Users don't compile any programs that link with the libraries
   provided in your EPICS binary packages, nor do they create any binary
   packages of their own that depend on your binary packages.

3. When you say "EPICS as binary packages," you mean only what has
   been in the past referred to as EPICS Base.  You do not mean EPICS
   modules.

4. You provide binary packages for EPICS as well as EPICS modules, and
   you build them all at once from the same source, and to upgrade, a
   user must upgrade all of the binary packages at once or none of them.

5. You provide one monolithic EPICS binary package that includes EPICS
   and all EPICS modules.

I'm not saying all of these will be true, but I think some or parts of
some of them must be true in order for it to work for you.  I'd be very
interested to hear how you have made things work using a binary package
management system.

Lewis

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System