Re: Re: ZEO vs NEO

I'd really like to hear what's been your take on asyncio now that
you're using it, Jim.

Regarding, Trollius, it's very close to beeing abandoned [1]. Victor
Stinner, the project founder and main committer has lost interest,
warned the community on Jan/25 and so far no one has taken over the
project:

[1] https://groups.google.com/forum/m/#!topic/python-tulip/mecVwhnVP0A

Anyway, I wouldn't see anything wrong with a Py3-only implementation
of ZEO, as long as it remains compatible with Py2 clients.

Cheers,

Luciano



On Sat, Feb 13, 2016 at 12:48 PM, Jim Fulton <[email protected]> wrote:
> I plan to finish the work I started on ZEO to port it to asyncio, as
> I'm pretty far along.  In addition to moving to a modern networking
> library (supporting SSL), this will allow me to greatly simplify the
> ZEO implementation, which is something I've wanted to do for some
> time.  (I expect that when I'm done, someone else will port the
> asyncio code to Trollius to get Python 2 support.)  As part of this
> I'll likely make ZEO a little bit more like NEO:
>
> - Per-object rather than global server commit locks, and a shorter
> window for commit locks.
>
> - Moving conflict resolution to clients by moving conflict-resolution
> handling into ZODB itself, probably through some pluggable mechanism
> outside the storage layers. This should allow some simplification of
> NEO and Relstorage.
>
> I also plan to devote a lot more time studying NEO.  Maybe someday NEO
> will replace ZEO, or maybe ZEO will serve applications that don't need
> multiple servers. We'll see.
>
> Jim
>
> On Tue, Feb 2, 2016 at 6:28 AM, Jim Fulton <[email protected]> wrote:
>> I really like the NEO architecture, at least at a high level.  I
>> haven't studied it in detail.  I've been planning to work toward a
>> similar architecture with ZEO.  A big stumbling block for me has been
>> NEO's GPL license.
>>
>> I was speaking with Jean-Paul Smets the other day and we discussed
>> adjusting NEO's license to make it more broadly usable.  He suggested
>> making the client LGPL.  Hypothetically, If NEO eventually replaced
>> ZEO and had an LGPL license, would the LGPL license present a problem
>> for anyone?
>>
>> If the LGPL isn't a problem, I'd put a lot more effort into looking at
>> NEO as a client-server architecture for ZODB, likely as a replacement
>> for ZEO.
>>
>> So does the LGPL present a problem to anyone?
>>
>> Jim
>>
>> --
>> Jim Fulton
>> http://jimfulton.info
>
>
>
> --
> Jim Fulton
> http://jimfulton.info
>
> --
> You received this message because you are subscribed to the Google Groups "zodb" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
> For more options, visit https://groups.google.com/d/optout.



-- 
Luciano Ramalho
|  Author of Fluent Python (O'Reilly, 2015)
|     http://shop.oreilly.com/product/0636920032519.do
|  Professor em: http://python.pro.br
|  Twitter: @ramalhoorg

-- 
You received this message because you are subscribed to the Google Groups "zodb" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
For more options, visit https://groups.google.com/d/optout.