On Tue, May 29, 2012 at 12:38 PM, Gervase Markham <[email protected]> wrote:
> On 28/05/12 11:18, Henri Sivonen wrote:
>> To avoid confusion, I think it's important to specify in the new
>> policy what versions of LGPL third-party code has to allow to be used
>> in order to be permissible in the Mozilla code base: for example,
>> whether an "or at your option any later version" choice is required
>> and whether it matters if the base version is 2.0, 2.1 or 3.
>
> OK...
>
> If I proposed "any of the above", would that be a problem?

I was thinking that specifying LGPL 2.0 or LGPL 2.1 without "or, at
your option, any later version" would have been a problem for using
the whole code base under the GPL, since using the whole code base
under the GPL will, in practice, mean GPLv3 once the code base has
become dependent enough on Apache-license code. However, it appears
the section 3 of LGPL 2.0 and 2.1 not only allows upgrading from LGPL
2.x to GPLv2 but automatically allows upgrading to a later version of
the GPL. So maybe LGPL 2.0 or 2.1 without "or, at your option, any
later version" isn't a problem.

Curiously, section 2 of LGPLv3 does not include an explicit
always-active version upgrade clause as part of the upgrade clause to
GPL. Obviously, that isn't a problem as long as version 3 is the
latest version of the GPL. It's unclear if that could be a problem in
the future.

So no, I don't, at present, see a concrete reason why "any of the
above" would be a problem.

>> Personally, I have occasionally silently wondered why Mozilla upgraded
>> to MPL 2.0 instead of moving to LGPLv3 with "or at your option any
>> later version" choice. If/when Mozilla allows LGPLed third-party code
>> in the code base, it would probably good to have some kind of
>> documentation that explains why first-party code needs to be under MPL
>> 2.0 in addition to being available under LGPLv3 (which it has ever
>> since the tri-license).
>
> There are several reasons; one of them is that companies wish to combine
> our stuff with proprietary code (a usage model which Mozilla has always
> enabled). This is harder with the LGPL because of the wider scope of its
> copyleft.

How realistic will it be to combine first-party Mozilla code with
proprietary code in an LGPL-incompatible way once Mozilla first-party
code becomes dependent on LGPLed third-party code? (This a question.
This is not an objection to allowing LGPL. The success of WebKit shows
that LGPL-compliance isn't a show-stopper burden for combining a Web
engine with proprietary code.)

-- 
Henri Sivonen
[email protected]
http://hsivonen.iki.fi/
_______________________________________________
legal mailing list
[email protected]
https://lists.mozilla.org/listinfo/legal

Reply via email to