Re: psql & readline & win32 - Mailing list pgsql-hackers

From Greg Stark
Subject Re: psql & readline & win32
Date
Msg-id 87fymn5act.fsf@stark.xeocode.com
Whole thread Raw
In response to Re: psql & readline & win32  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
Tom Lane <tgl@sss.pgh.pa.us> writes:

> > To clearify, I meant the psql binary becomes GPL.
> 
> There is no such thing as "the binary becomes GPL".  GPL applies to
> the source code.

That's an odd thing to say. The binary is as much covered by copyright as the
source and can't be distributed without satisfying the requirements of the
license that covers it. The GPL requirements mean you can't distribute a
binary that depends on readline without including the corresponding source
code.

I'm not sure that's really an onerous requirement. It just means if you're a
commercial vendor selling a binary-only version of Postgres you can't link
your binary-only version against readline and then distribute it. Which should
be pretty obvious anyways. 

(The exception Tom points out might even make it legal to distribute a Linux
compile of Postgres linked against readline since most Linux distributions
include readline. That wasn't true when that exception was written though so
you may want to check with your lawyer about that.)

I think people are mixing this stuff up with the less obvious claim about
programs like postgres being deemed "derivative works" of libraries like
readline because they "depend" on them. Postgres doesn't really depend in any
real sense on readline so I can't see that argument working in this case
anyways. If there was some GPLed library that Postgres couldn't work usefully
without then there might be a real need for a non-GPL'd version of that
library.

-- 
greg



pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: Using stats_command_string for xact statistics
Next
From: Andrew Dunstan
Date:
Subject: Re: Why don't we allow DNS names in pg_hba.conf?