Am 27.11.2013 um 14:05 schrieb Sabine Knöfel <sabine.knoe...@gmail.com>:
> Hi Norbert, > > thank you very much for the clarification. > > I had to read it several times but now I understand and I know what to > change in my model for this (something with references to objects which have > been deleted by the user). > Yep, it _is_ a horrible explanation but the only one I can come up in a short timeframe :) Making it easily readably and clear would talk some more time. > Perhaps your answer is also interesting for the documentation of Johan. > Uh? I’ve never seen a documentation about Johan. That is a nice idea! I was always thinking about how to use him! Norbert > Greets > Sabine > > > On Wed, Nov 27, 2013 at 1:44 PM, Norbert Hartl [via Smalltalk] <[hidden > email]> wrote: > > Am 27.11.2013 um 13:20 schrieb Sabine Knöfel <[hidden email]>: > > > Hi Esteban, all, > > > > I work with mongo daily an it works fine, I am very happy with it. > > > > As you told me, with > >>> VORepository current reset. > > I can force re loading all objects from database and resetting the cache > > completely. > > This is helpful for development eg. after changing the magritte > > descriptions. > > > Indeed. You need to flush/reset the repository in order to have new > descriptions to take effect. > > > My question: > > Is there a possibility to make a query and tell voyage that THIS query > > should be done within the database and NOT within the cached objets? And > > that all objects and the child objects from this query are loaded freshly > > from database into the cache. > > > All queries go to the database. The cache is just used at resolve time. And > it is necessary to have identical objects. Meaning you query the database > directly and then objects are materialized. If the object is already in the > cache the cached one is returned. Otherwise you would lose identity because > having two requests containing the same object as a result would lead to two > objects instead of one. You could only load trees instead of a graph. > > > My concrete situation: > > I want that if the user logs in, his objects are loaded from database and > > NOT from cache. This means the person, his trips etc. > > > > And the question coming along with this: > > How long does the cache keeps objects/when are they reseted (except the > >>> VORepository current reset)? > > If I would never take a new image and never make a reset, would all objects > > remain in the image (and the database objects wold never be read)? In this > > case what about the size of the image? > > > I’m not sure on this one. I think the cache does not clean anything. The > thing about being able to load a graph I wrote above has also the constraint > that all loaded objects that are still participating in the active graph need > to be in the cache. But then the cache is a weak dictionary meaning that all > objects leaving the active graph are removed from the cache (well at GC time, > I think) > > Norbert > > > Regards > > Sabine > > > > > > > > -- > > View this message in context: > > http://forum.world.st/Mongo-cache-vs-database-objects-tp4725554.html > > Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com. > > > > > > > If you reply to this email, your message will be added to the discussion > below: > http://forum.world.st/Mongo-cache-vs-database-objects-tp4725554p4725577.html > To start a new topic under Pharo Smalltalk Users, email [hidden email] > To unsubscribe from Mongo cache vs database objects, click here. > NAML > > > View this message in context: Re: Mongo cache vs database objects > Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.
