Re: advice on MountedFileStorage changes.
Shane Hathaway wrote I looked some more at ZClasses and such today, hoping to solve the problem with mounted databases. Guess what! It turned out to be a simple matter.
Excellent! I've added support in CVS, and I'll do a new release this evening.
The latest addition to Mount.py provides the ability for a MountPoint to get its ZClass definitions from the root database rather than the mounted database. Anthony, I think you'll want to make "classDefsFromRoot" a checkbox available when adding a new mount point.
Done. Cool, it looks like it even handles recursive mounts.
Depending on the application, this can be the more sensible (and more manageable) approach. For example, you may want to have a large collection of little FileStorages, all using the same ZClass definitions. You certainly wouldn't want to make a copy of all ZClasses in each FileStorage.
Exactly. For me, one of the big things about the MountedFileStorage is the ability to split out the data for various sections of the tree - I still want to be able to make ZClass instances, but I don't want to have to manage copying all products all over the place.
In doing this I realized, and Jim suggested, that there are other situations where you'd really want ZClass definitions to come from the mounted database. This is a larger issue. We will first need to provide a way to manage the databases. Then there will be a need to load/unload products dynamically. I don't think we want to try to finish that for 2.2.0.
The next MountedFS release will have a teeny tiny Control_Panel like thing in the root of the mountpoint. I'll try putting a Products folder in it. Anthony
participants (1)
-
Anthony Baxter