[PATCH] https: support tls sni (server name indication) for https urls (issue3090)

Augie Fackler raf at durin42.com
Wed Sep 18 07:43:40 CDT 2013


On Mon, Sep 9, 2013 at 4:37 PM, Alex Orange <crazycasta at gmail.com> wrote:

> Well, it has taken me a little bit to think about this. I am a strong
> proponent of GPL and AGPL. However, because I have found Mercurial so
> useful in the past, I have decided to relicense the current version of my
> ssl_sni code (which I believe satisfies all the needs of Mercurial) under
> GPLv2 or any later version (as I understand Mercurial is licensed).
>

Thanks.


> I will also provide any support as far as bug fixes go to make sure that
> it is a quality product. To be clear however, the ssl_sni package on pypi
> will remain AGPLv3, and any further changes I make to it will remain
> AGPLv3.
>

Honestly, I'm not sure I'd put too much effort into an AGPL library -
there's just not much point, given that relatively few folks can embed
AGPL-licensed anything into their address space.

(This isn't me trying to convince you to not AGPL this, I'm just trying to
point out the relative absurdity of a virally licensed library that forces
any of its clients to be AGPL as well.)


> I will repackage the code and make the license clear as GPLv2 or later and
> send a new patch to the list.
>

Thanks, I look forward to seeing the patch. I think at this point, you
should just go ahead and ignore httpclient entirely, and I'll figure out
what the fixed interface for httpclient should look like upon seeing your
patch (it probably just needs an ssl_factory keyword argument, but I'll
withhold that decision until I see your patch).


>
>
> On Tue, Sep 3, 2013 at 4:59 PM, Matt Mackall <mpm at selenic.com> wrote:
>
>> On Tue, 2013-09-03 at 14:50 -0400, Augie Fackler wrote:
>> > On Thu, Aug 29, 2013 at 11:49:13AM -0600, Alex Orange wrote:
>> > > # HG changeset patch
>> > > # User Alex Orange <crazycasta at gmail.com>
>> > > # Date 1377769698 21600
>> > > # Node ID c9e262317f32cf3253d83384f3232577fb565675
>> > > # Parent  d4a0055af149cdea20b3136b66cae8a24b2e2a98
>> > > https: support tls sni (server name indication) for https urls
>> (issue3090)
>> >
>> > This chunk of the solution looks fine. Sadly, since the ssl_sni
>> > package is licensed under the AGPL, I can't use it (and so haven't
>> > looked at it) due to our policies at work.
>> >
>> > Can I persuade you to view this package as enough of a public good
>> > that it be licensed under some super-permissive (MIT? Apache2?)
>> > license, rather than the hyper-viral AGPL?
>>
>> Agreed. Generally speaking, the Mercurial project can't take any
>> contributions that use or link to anything that is not GPL _v2_ or
>> weaker.
>>
>> --
>> Mathematics is the supreme nostalgia of our time.
>>
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://selenic.com/pipermail/mercurial-devel/attachments/20130918/85d77ed1/attachment.html>


More information about the Mercurial-devel mailing list