ZODB, Databases, and Threads

Ixokai noone at nowhere.org
Wed Apr 11 00:24:15 EDT 2001


Hello,

    I'm working on a project using Stackless Python so that I can have quite
a few threads of execution running at the same time without the overhead of
creating and destroying them, since the above will all be done possibly
frequently.

    Now. I also want an database, which will be either Object-oriented, or
I'll just teach my classes to store themselves (via picke?) in the database.

    I was origionally going to use ZODB, but after doing some looking into,
it seems as though it might not be appropriate. You see, in my application,
it is ABSOLUTELY ESSENTIAL that every thread access the SAME data at any one
moment in time; none of this transaction bit. I was going to do a 'many
readers, one writer' bit with a lock to make it so only one thread can
write, but it sounded like Zope with its get_transaction() bit was actually
geared to make this difficult.

    Is there another option? I' rather like the idea of 'versioning' and
'undo' capabilities, so I was looking at bsddb3 from Sleepycat and wondering
about just writing some persistance code on top of it.

    That leads me to my second question -- how the heck does Zope go about
managing persistance while still letting people override __getattr__ and
__setattr__? :)

Thanks.

--S

(replacae 'NOSPAM' with 'seraph' to respond in email)




-----= Posted via Newsfeeds.Com, Uncensored Usenet News =-----
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
-----==  Over 80,000 Newsgroups - 16 Different Servers! =-----



More information about the Python-list mailing list