If one looks at the documentation at the hsqldb site , it is recommended
that the server mode be used during development and only after one is
certain the wrapper program is functioning properly that the move to the
embedded model be made. This makes perfect sense when you think about
it, but just the opposite is what is going in the Base environment. A
user fires up the Base module and starts to build the database - telling
them, to go install the hsqldb server, build against it, and then move
the database back to the embedded model just is not a viable choice.
My starting assumption is this - When the choice was made to use an in
memory model for the native data engine there had to have been some
thought about inherint risk of data loss. I therefore assume that some
design decisions, or reasons, where put forward regarding mitigation of
this risk.
It has become quite important, for me, to try and get a good handle on
just when data is being written to permanent storage during runtime,
using the embedded model. I know that some changes where made to the
code between 2.0.0 and 2.0.1, don't know if any more are being included
in 2.0.2.
Would it be possible for someone to give me a brief -bullet item list of
when, or what events trigger, this write back to permanent storage. I
now have a copy of the 2.0.1 source code on my machine, so if you need
to refer me to specific modules for review as the best way to get this
information that is acceptable. I would then most likely come back with
specific questions from my review of the code.
Thanks very much for any assistance
Andrew 'Drew' Jensen
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
