-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Jürgen kartnaller wrote:
I just run into a with transaction manager and synchronizers.
If transaction.get implicitily creates a new transaction the synchronizers are not called. This is not problem as long as no savepoints are used. If savepoints are used the DataManagers created by synchronizers are not present. Because savepoints are requested from the DataManagers it doesn't work.
Can you elaborate on how your patch interacts with this comment (in transaction._manager)? # Call the ISynchronizer newTransaction() method on every element of # WeakSet synchs. # A transaction manager needs to do this whenever begin() is called. # Since it would be good if tm.get() returned the new transaction while # newTransaction() is running, calling this has to be delayed until # after # the transaction manager has done whatever it needs to do to make its # get() return the new txn. In other words, is there any cases there TM.get() will be called when it is "too early" to call '_new_transaction'?
I already created a branch with a fix (transaction/branches/synchfix).
If one with more knowledge about the transaction manager can approve my change I will add tests and merge the change into trunk.
It would help to see the tests with the rest of the patch: they often help illustrate the motivation. Tres. - -- =================================================================== Tres Seaver +1 540-429-0999 tseaver@palladion.com Palladion Software "Excellence by Design" http://palladion.com -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFI15wu+gerLs4ltQ4RAiA8AJ9hLInfO/5c5633rgqY784YdBjsTgCgoMjM gAA7c5VH4jhy+wTVt+uXztk= =rMB8 -----END PGP SIGNATURE-----