Re: [announce] NEO 1.0 - scalable and redundant storage for ZODB
Julien Muchembled <[email protected]>
| Newsgroups | gmane.comp.web.zope.zodb,gmane.comp.web.zope.devel |
|---|---|
| Message-ID | <[email protected]> |
Le 08/29/12 15:54, Leonardo Santagada a écrit : > On Mon, Aug 27, 2012 at 9:37 AM, Vincent Pelletier <[email protected]> wrote: >> We've just tagged the 1.0 NEO release. > > any plans to support pypy in NEO? We're looking at pypy from time to time but there are at least 3 reasons preventing us to use it for NEO: - sqlite3 is still significantly slower on pypy so it could not be used for SQLite backend. I've just tested the example on https://bugs.pypy.org/issue1070 with version 1.9.0 and it's more than 2 times slower and what is suggested on https://bugs.pypy.org/msg3088 does not help. - MySQLdb is C module. We'd have to use another library for MySQL backend but that should be quick to implement. - For the client part, as you say, it depends on ZODB. So it should be possible to use pypy with admin/master/storage nodes and maybe get better performance with MySQL backend. > Once ZODB start to run on pypy it > would be cool to have relstorage/zeo or neo running on it as well. > Julien _______________________________________________ For more information about ZODB, see http://zodb.org/ ZODB-Dev mailing list - [email protected] https://mail.zope.org/mailman/listinfo/zodb-dev