Subject: Re: EROS license
From: Ian Lance Taylor <>
Date: 26 Jun 1999 12:16:33 -0400

   Date: Sat, 26 Jun 1999 12:02:04 -0400

   > Where did you get the idea that Cygnus has ever done this?

   I was the recipient of the distribution.

   The context was the SPARC-v9 processor, which at the time had not yet been
   released.  Information on this processor was only available under NDA, and in
   consequence, the compiler could not be distributed except to parties who had
   signed the NDA.  Sun, HaL, and I think some others contracted with Cygnus to
   have gcc ported.

   Strictly speaking, the constraint that the recipient of this working compiler
   (i.e. Sun, HaL, etc.) must have signed an NDA to receive the compiler violates
   the GPL.  The relevant GPL text is:

	You must cause any work that you distribute or
	publish, that in whole or in part contains or is
	derived from the Program or any part thereof,
	to be licensed as a whole at no charge to all
	third parties **under the terms of this License. **

   Those last words (emphasis mine) specifically preclude additional terms, which
   would preclude an NDA restriction.  The distribution of the prerelease compiler
   by Cygnus to HaL was therefore a technical violation of the license.

This was admittedly before my time at Cygnus.

You said that Cygnus made a binary only release.  That is what I was
disputing.  Did you actually get that software directly from Cygnus?
In the six years I was at Cygnus, we were always careful to include
the sources with every binary distribution we made.

Are you sure that, for example, you didn't get the binaries from Sun?
Are you sure that Cygnus imposed any restrictions on you, rather than,
say, Sun imposing restrictions on you?

If Sun made a binary only release, or if they imposed restrictions on
you such as prohibiting you from redistributing the compiler, then I
agree that Sun violated the GPL.  I am just extremely surprised that
you are claiming that Cygnus violated the GPL.