Hi!
Brent Pinkney wrote:
Hi Milan,
Ther will be a lot of documentation forthcoming on Magma soon(-ish).
Cool! :)
I would suggest the Lava test cases as a decent place to start.
I would also ping this list for help.
Brent
As noted Gjallar uses Magma but definitely not in the "easiest manner". :)
Since Gjallar uses "reified transactions" in the form of a class hierarchy with one subclass for each "kind" of database change - and then instantiates such a class, feeds it with the proper data, and then tell it to "execute" on the model, finally putting it away for historical reasons in an ever growing MagmaCollection - it definitely makes it hard to "find" the use of Magma in Gjallar.
It is really not a good example of how to use an OODB/Magma in the "vanilla" case.
Having said that Gjallar is still an interesting example since we use both strategies, indexing and quite a complex object model - all working quite smoothly.
For example, we have a customer running a Gjallar instance since january without any issues - the VM has not even been restarted since then.
regards, Göran