the bigger issue is how should a cache implementation be integrated
with SQLAlchemy in the first place. I dont think caching is so
simple; for example, I dont just want to cache an object based on
its ID. I want to cache the results of all kinds of queries which do
expensive joins and sorts as well. so, do we cache the results of
all SQL statements as well ? its not always appropriate, and there
are a lot of variables that might change on a query-by-query basis.
So, its not clear to me how a cache would be "integrated" with
SA...because the usage patterns would vary greatly, it still seems
like a layer that should ride on top of SA and be specific to a
particular application. If someone were to contribute such a layer
and it became obvious that there are two or three patterns that are
extremely common and widely useful, that would begin to formulate how
such a layer could be more integrated. I dont yet see the value in
a straight "id" cache, which is very easy to implement by oneself,
and were it included as core functionality would result in a whole
new crop of user errors and confusion that then get reported as bugs
(i.e. stale objects, inconsistent results as some objects get flushed
from the cache others dont, etc.).
generally, I am trying to keep SQLAlchemy as simple as possible, and
I am avoiding any feature which solves only a small portion of a
particular prolbme space...it has to be a total killer feature
otherwise its not worth the support. if you can propose a total
"killer" version of it, im definitely interested.
On a related note, if you are interested in Memcached, there is a
container API already being looked at by the Pylons and Turbogears
crew which I wrote for Myghty which provides "switchable" backends to
flat files, DBM files, dictionaries, and memcached, originally
written to replace Cache::Cache which is used with HTML::Mason. If
I were to implement caching in SQLAlchemy, I might very well use that
API (or an enhanced version of it).
On Mar 14, 2006, at 6:39 AM, Florian Boesch wrote:
Hi,
I just stumbled upon memcached, which seems to be one of the finer
implementations of distributed tuple spaces. It is used for caching
objects.
Any idea if things like this could be integrated with sqlalchemy.
Cheers,
Florian
-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting
language
that extends applications into web and mobile media. Attend the
live webcast
and join the prime developer group breaking into this new coding
territory!
http://sel.as-us.falkag.net/sel?
cmd=lnk&kid=110944&bid=241720&dat=121642
_______________________________________________
Sqlalchemy-users mailing list
Sqlalchemy-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sqlalchemy-users
-------------------------------------------------------
This SF.Net email is sponsored by xPML, a groundbreaking scripting language
that extends applications into web and mobile media. Attend the live webcast
and join the prime developer group breaking into this new coding territory!
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642
_______________________________________________
Sqlalchemy-users mailing list
Sqlalchemy-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/sqlalchemy-users
