[Monotone-devel] Re: optimized SHA1

[Nathaniel Smith]

On Sat, Aug 05, 2006 at 11:38:55AM +0200, Richard Levitte - VMS Whacker wrote:
> In message <address@hidden> on Sat, 5 Aug 2006 00:40:17 -0700, Nathaniel 
> Smith <address@hidden> said:
> 
> njs> Richard, I'd appreciate any feedback or tips you have before
> njs> going forward on this -- you know the OpenSSL folks and I don't,
> njs> and I don't want to step on any toes or accidentally run into
> njs> some history I don't know about :-).
> 
> Simply ask if we can grant you permission to use specific files from
> OpenSSL under a different license.  There might be some harsh feelings
> about the GPL, but if you ask for a relaxed variant of the OpenSSL
> license, such as BSD3p or BSD2p :-), I don't see a reason not to.

Right, that's what I figured.

> I'll certainly vote FOR for such a move, and I don't think it would
> come as a surprise, as I've talked about monotone with the rest of the
> guys already.  In reality, it's up to the authors, not the entire
> OpenSSL group.

Nod -- that's what the law says, if nothing else :-).  (Unless you
guys actually do copyright assignments, but I'm pretty sure that's
not the case.  Umm... I guess if it is I need to know, though :-).)

It actually occurs to me that it might be extra-cool if notices were
added to those files in OpenSSL CVS, saying "NB, this file is licensed
under a more permissive license than OpenSSL as a whole; you are free
to disaggregate it and redistribute it under the terms of a 2-clause
BSD license" (or whatever).  This would be handy because it would mean
we wouldn't have to repeat this shenanigan every time someone
committed fixes to those files; as long as they left the license
notice there, it could be assumed they were okay with their fixes
being re-used elsewhere without the advertising clause.

> Hmm, maybe I could even drive the question myself.

Well, you know how eager I am to do everything by myself without any
help, since I have so much empty time I need to fill up somehow...

...but, well, I'll make an exception in this case; feel free if you
want to ;-).

> You know, there's another way as well, and it's to have a note in the
> monotone code that gives permission to link the code with OpenSSL.  I
> think it's been described more than once how to do this.

Right -- but then _we're_ GPL-incompatible.  Aside from being a
tremendous hassle (there are 56 names in AUTHORS, and in principle all
would need to relicense their contributions; the lru cache code
Graydon downloaded from somewhere is GPLed; ...), this falls into that
class of "workarounds that make me very queasy".  The whole point of
the FSF and the GPL is to create an enclave wherein we don't have to
deal with this copyright junk.  I'm uncomfortable with expanding
ghettos of incompatibility.  (...Which reminds me that we _still_ need
to do the relicensing dance with the manual.  Sigh.)

The advertising clause is also reasonably noxious in its own
right :-).

> njs> (Also, do you by chance understand the tangle of code that calls
> njs> into the asm cores?  I'm sure we could work it out, but it sure
> njs> isn't as easy to read as the clean self-contained stuff like the
> njs> git asm kernels, e.g:
> njs>   http://www.kernel.org/git/?p=git/git.git;a=tree;f=arm
> njs> )
> 
> I think I can understand it if I put my head to it.  At this point,
> though, Andy is out assembly master :-)

Yeah, I sort of got that impression looking at the checkin logs :-).

Thanks,
-- Nathaniel

-- 
Linguistics is arguably the most hotly contested property in the academic
realm. It is soaked with the blood of poets, theologians, philosophers,
philologists, psychologists, biologists and neurologists, along with
whatever blood can be got out of grammarians. - Russ Rymer