HI!
It seems TXN support has landed in git master. Is it ready for interop testing?
Ciao, Michael.
Michael Ströder wrote:
HI!
It seems TXN support has landed in git master. Is it ready for interop testing?
Yes, go for it.
Howard Chu wrote:
Michael Ströder wrote:
It seems TXN support has landed in git master. Is it ready for interop testing?
Yes, go for it.
More questions:
In case of using Assertion Control: At which state is the assertion tested? During TXN settlement?
If an update request is sent by the client with a TXN identifier is that passed through the whole overlay stack?
More specific:
Is the result code of the update operation not success(0) if it would have failed due to violated constraint (slapo-unique, slapo-constraint, slapo-refint etc.)? Or is this overlay processing deferred until TXN settlement?
What about slapo-accesslog? Is the update operation immediately written to the accesslog DB? Or is it deferred until TXN settlement?
Hmm, many subtle issues might pop up for one or the other...
Ciao, Michael.
Michael Ströder wrote:
Howard Chu wrote:
Michael Ströder wrote:
It seems TXN support has landed in git master. Is it ready for interop testing?
Yes, go for it.
More questions:
In case of using Assertion Control: At which state is the assertion tested? During TXN settlement?
During TXN settlement.
If an update request is sent by the client with a TXN identifier is that passed through the whole overlay stack?
The control is passed thru the Operation struct.
More specific:
Is the result code of the update operation not success(0) if it would have failed due to violated constraint (slapo-unique, slapo-constraint, slapo-refint etc.)? Or is this overlay processing deferred until TXN settlement?
Section 3.3. All processing is deferred until settlement.
What about slapo-accesslog? Is the update operation immediately written to the accesslog DB? Or is it deferred until TXN settlement?
Same answer.
Hmm, many subtle issues might pop up for one or the other...
Ciao, Michael.
-
Howard Chu -
Michael Ströder