Re: [HACKERS] Readline use in trouble? - Mailing list pgsql-hackers

Bruce Momjian <maillist@candle.pha.pa.us> writes:
> Here is something I read as part of the Alladin Ghostscript 6.0 beta
> release.  I must admit I don't understand the logic of the issue.  It
> seems the issue is that you can link non-GPL to GPL libraries, but you
> can't distribute the result.   Maybe it doesn't apply to us because we
> don't copyright our code.

Huh?  We certainly do --- or have you missed that* Copyright (c) 1994, Regents of the University of California
that's plastered across all the source files?

The GPL does restrict the conditions under which GPL'd code can be
distributed; in particular it can't be distributed as part of a program
that is not all GPL'd (more or less --- I have not read the terms lately).
So, because we use BSD license rather than GNU, we cannot *include in
our distribution* any library that is under GPL.

Any end user who does not intend to redistribute the result can
certainly obtain our distribution and readline and build them together.
So it's no issue for source distributions, but I wonder about RPMs.
Our RPMs do not include the actual libreadline file, do they?

>    Even though the GNU License (GPL) allows linking GPL'ed code (such as
>    the GNU readline library package) with non-GPL'ed code (such as all
>    the rest of Ghostscript) if one doesn't distribute the result, the
>    Free Software Foundation, creators of the GPL, have told us that in
>    their opinion, the GPL forbids distributing non-GPL'ed code that is
>    merely intended to be linked with GPL'ed code.

As stated, this is ridiculous on its face.  The FSF has no possible
right to prevent the distribution of software that they didn't write
and that doesn't fall under the GPL.

Although I haven't been paying close attention to the Ghostscript
situation, I suspect that the real story is either that the readline
interface code that someone contributed to Ghostscript was contributed
with GPL terms already attached to it, or that Aladdin is concerned
about being able to distribute full-featured precompiled binaries of
Ghostscript.  (BTW, Peter Deutsch has a history of forcing the issue
when he thinks that someone else is being unreasonable, and I suspect
that he's deliberately overreacting in hopes of making FSF change
their position.)

Anyway, this sort of thing is why it's a bad idea to accept any GPL'd
code into Postgres --- the GPL does not play nice with other licenses.
I think the FSF is not doing the free software movement any service
with this foolishness, but they're entitled to distribute their code
with any terms they want, of course.

My inclination is to ignore the issue until and unless we hear a
complaint from the libreadline authors --- and if we do, we yank all
trace of readline support from psql.  End of story.
        regards, tom lane