Re: Discussion about why GNU/Linux system upgrades cause old programs to break

Richard Stallman <[email protected]> writes:

> [[[ To any NSA and FBI agents reading my email: please consider    ]]]
> [[[ whether defending the US Constitution against all enemies,     ]]]
> [[[ foreign or domestic, requires you to follow Snowden's example. ]]]
>
>   > Alternatively, a program should be built in a way that it links to a
>   > generic libFOO.so, not to a version-specific libFOO-nn.so.0, so that
>   > upgrading will redirect the generic link to the actual library, and
>   > programs will still work (provided that the new library is
>   > ABI-compatible with the old one).
>
> Why/when do people link programs with the version-specific form
> library?

I've never seen that happen.  sonames, which are used for identifying
libraries, typically don't change unless a breaking change happens.

> Do you think people misguidedly request that?
> Do our tools sometimes lead people to do that by default?
> Can you suggest changes we could make to avoid such occurrences?
>
>   > Not removed yet.  But, having read the relevant discussions, I won't
>   > be surprised if it will become an error (thus the build will be
>   > broken), and if the option to make it compile without warnings will be
>   > removed at some point.
>
> We should _never_ go to that extreme.

Someone needs to step up to define what "compatibility" means, in that
case.

It is hard to stay compatible with a language with no known semantics.

The boat mostly doesn't get rocked much at the moment, but clearly the
rocking is too much for some.

>   > I always thought that in the GNU project we don't blindly follow The
>   > Powers That Be out there when they decide to remove or EOL something.
>   > Instead, we make our own decisions, based, in no small measure, the
>   > benefits and disadvantages for our users.
>
> Yes exactly -- and this means we do not adopt unquestionable faith 
> in the decisions of a standards committee.

I think you're vastly overestimating the differences in what our
conclusion would be and what such a committee might conclude.

Especially now, given that GNU decisions inform and staff sit on said
committees.

In addition, IME, current differences to existing standards are a
greater source of confusion than benefit.  Introducing new differences
is likely to be even worse.

Luckily, I doubt that'll be something to frequently consider.
-- 
Arsen Arsenović

-----BEGIN PGP SIGNATURE-----

iQECBAEWCgCqFiEE/uKz0RP8AKMWLWBhUsKUMB6ixJMFAmp+KkQbFIAAAAAABAAO
bWFudTIsMi41KzEuMTIsMiwyXxSAAAAAAC4AKGlzc3Vlci1mcHJAbm90YXRpb25z
Lm9wZW5wZ3AuZmlmdGhob3JzZW1hbi5uZXRGRUUyQjNEMTEzRkMwMEEzMTYyRDYw
NjE1MkMyOTQzMDFFQTJDNDkzEBxhcnNlbkBhYXJzZW4ubWUACgkQUsKUMB6ixJMk
6wD+KDVbTUkX7PB3IcmN3CHWTiRmKjCZEcdfT4jhZJKxXx0A/35aBkB0uB5G9xNT
IGY30oeaHquZV+U0kuT2Rh+qHhoH
=j/2j
-----END PGP SIGNATURE-----