git-svn-id: https://svn.erp5.org/repos/neo/trunk@2000 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1999 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1998 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1997 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1996 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1995 71dcc9de-d417-0410-9af5-da40c76e7ee4

Typical case:

- Handler H1 is set
- Send request R1
- Request switch to handler H2 (delayed)
- Request switch to handler H3 (delayed)
- Receive answer A1
- H2 must be discarded and H3 applied on the connection

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1994 71dcc9de-d417-0410-9af5-da40c76e7ee4

- Store None values as NULL column values
- Raise KeyError if the configuration entry is not found

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1993 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1992 71dcc9de-d417-0410-9af5-da40c76e7ee4

Remove admin's app.ptid attribute, use pt.load() and pt.update() instead
of copy/paste the code.

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1991 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1990 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1989 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1988 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1987 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1986 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1985 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1984 71dcc9de-d417-0410-9af5-da40c76e7ee4

This fix a race condition when switching from recovery to verification
steps where NotifyLastOID packet could be handled by verification handler
instead of initialisation (storage side).

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1983 71dcc9de-d417-0410-9af5-da40c76e7ee4

This packet handler should be dropped but it's break some functional tests,
may be due to a race condition when switching from recovery to verification.

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1982 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1981 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1980 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1979 71dcc9de-d417-0410-9af5-da40c76e7ee4

This mimics what FileStorage uses (file offsets) but in a relational manner.
This offloads decision of the ability to undo a transaction to storages,
avoiding 3 data loads for each object in the transaction at client side.
This also makes Neo refuse to undo transactions where object data would happen
to be equal between current value and undone value.
Finally, this is required to make database pack work properly (namely, it
prevents loosing objects which are orphans at pack TID, but are reachable
after it thanks to a transactional undo).

Also, extend storage's transaction manager so database adapter can fetch data
already sent by client in the same transaction, so it can undo multiple
transactions at once. Requires making object lock re-entrant (done in this
commit).

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1978 71dcc9de-d417-0410-9af5-da40c76e7ee4

OID 0 is a very valid oid. It's often used for the most important object of
the database: the root object. Using 0xffffffffffffffff leaves some space.

TID 0 is a very valid tid, for transactions back in unix year 0... or
ZODB tests. Using 0xffffffffffffffff leaves some space.

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1977 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1976 71dcc9de-d417-0410-9af5-da40c76e7ee4

Test case (added):
- Start a cluster with two storages
- Restart the whole cluster except one storage
- Allow recover from existing storage
- The admin crash when getting the new partition table because it doesn't
known the second (down) node.

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1975 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1974 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1973 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1972 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1971 71dcc9de-d417-0410-9af5-da40c76e7ee4

BLOB is 2^16 bytes
MEDIUMBLOB is 2^24 bytes
LONGBLOB is 2^32 bytes

Choose LONGBLOB to store bigger objects.

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1970 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1969 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1968 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1967 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1966 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1965 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1964 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1963 71dcc9de-d417-0410-9af5-da40c76e7ee4

ReadBuffer join received strings only when all requested data is available.
This avoid many useless data copies in case of big packets. The gain factor
is about 50x for a 25MB packet.

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1962 71dcc9de-d417-0410-9af5-da40c76e7ee4

git-svn-id: https://svn.erp5.org/repos/neo/trunk@1961 71dcc9de-d417-0410-9af5-da40c76e7ee4