Hi Tim,
I have only modified the transaction package, not anything else and will not make any changes to the Id package, of course. I have checked in my stuff and class FileSequence in package org.apache.commons.transaction.file is the interesting one. It is pretty mature and I do not think it will change until the first release of the transaction package. Currently, it stores long only, but it should be very easy to adapt it to characters or anything else.
Looking forward to suggestions how efforts can be merged.
Cheers, Oliver
Tim Reilly wrote:
Hi Oliver,
[Phil Steitz wrote:]
Sounds like a good addition / enhancement to the serial
generators to me. You are correct. Only the UUID and SessionIdGenerators will be unique across instances.
Sounds good to me as well. Look forward to it.
[Oliver Zeigermann wrote:] I will check the stuff in next week. Let's see how we can combine efforts then.
One thought, the actually persistence mechanism should be a choice/pluggable IMO. I included the ReadWriteFileSystemImpl implementation in the o.a.c.id.uuid.state package but wanted to leave the interface open for other choices (jdbc, jndi, etc.) It may be possible to refactor that to use your persistence mechanism? Since explicit I/O is forbidden in some containers that may use id, I felt it must be a pluggable (hence o.a.c.id.uuid.state.. ReadOnlyResourceImpl, and InMemoryOnlyImpl`s as well.)
Thanks, -TR
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
