[ZODB-Dev] IIBTree.multiunion and list comprehensions

[Tim Peters]

[Tim Peters]
>> Whenever Jeremy has done a dump of an .fs file from the field, we've
>> usually found that most BTrees in it are tiny (no more than a few
>> dozen entries).

[Christian Robottom Reis]
> That's quite surprising. Do you mean that even when looking at
> ZCatalogs for large collections of pages/data you don't find massive
> BTrees?

Nope, but that's reading a lot into what I said.  When Jeremy does a dump of
an .fs file from the field, we generally have no idea what's in it, and
don't really care -- he gets one to begin with because it's damaged in some
way, and we're only mining it for clues about the source of the damage then.
Most BTrees in all .fs files we've looked at are in fact small.  There's no
distinction here made between BTrees used for catalogs and BTrees used for
all other purposes, and no claim either that *all* BTrees are small.

> I wonder what's the canonical `extreme ZODB.BTrees application', if
> such a thing exists.

I indeed expect that catalogs build the biggest BTrees normally seen in
practice.