[ZODB-Dev] Preventing a folder from being packed
Shane Hathaway
shane at zope.com
Thu May 15 15:21:58 EDT 2003
Jeremy Hylton wrote:
> On Thu, 2003-05-15 at 13:48, sean.upton at uniontrib.com wrote:
>
>>Forgive my ignorance, but wouldn't this be a bit of work? Since ZODB
>>essentially just persists objects and references, I wonder how much of a
>>mismatch and additional complexity the Zope-level notion of
>>folders/containment hierarchies would add to the storage. How smart should
>>the storage be? I like the per-class adjustment of history, because this
>>maps well in and out of Zope, but wouldn't notions of containment be harder
>>to preserve?
>
>
> Here's an unbaked idea.
>
> I think it could work on a per-transaction basis by using transaction
> metadata that was written by the application and read during pack. This
> would keep more revisions than necessary in many cases; a txn modifies
> two objects, one that it wants revisions for and one that it doesn't.
> The practicality of the approach depends in part on how likely an
> application is to modify a mix of objects in a single transaction.
Zope applications frequently write to the catalog at the same time. You
wouldn't want the catalog changes to be considered as permanent as the
folder contents. So much would be permanent that you'd be back to
square one.
I'm with Sean. It seems like getting this right would be difficult,
especially if you want all descendants of a folder to be permanent.
Mounted databases already address 90% of the need. IMHO the only
remaining 10% is making it easy to move an object into a new mounted
database.
Shane
More information about the ZODB-Dev
mailing list