Hi everybody,
When a story gets to have lots of versions (say, 3000+), and has also
had lots of stuff done to it (republishes, saves, edits and so on),
cancelling a checkout on that story gets agonizingly slow.
Greg had a look at why this was, and concluded that it was because
Bricolage was taking a long time to look through the story's event log,
so that it could be returned to the location (desk, shelf, etc.) where
it was before the checkout happened.
This makes sense for stories with short event histories. But asking the
database to sort a million event entries by date just takes a long time
no matter what.
So we were wondering: might it make sense to just store the pre-checkout
state someplace dedicated to the task, so it could be quickly looked up,
rather than having to dig it out of the event log?
Thoughts?
Thanks,
Bret
--
Bret Dawson
Producer
Pectopah Productions Inc.
(416) 895-7635
bret@pectopah.com
www.pectopah.com
When a story gets to have lots of versions (say, 3000+), and has also
had lots of stuff done to it (republishes, saves, edits and so on),
cancelling a checkout on that story gets agonizingly slow.
Greg had a look at why this was, and concluded that it was because
Bricolage was taking a long time to look through the story's event log,
so that it could be returned to the location (desk, shelf, etc.) where
it was before the checkout happened.
This makes sense for stories with short event histories. But asking the
database to sort a million event entries by date just takes a long time
no matter what.
So we were wondering: might it make sense to just store the pre-checkout
state someplace dedicated to the task, so it could be quickly looked up,
rather than having to dig it out of the event log?
Thoughts?
Thanks,
Bret
--
Bret Dawson
Producer
Pectopah Productions Inc.
(416) 895-7635
bret@pectopah.com
www.pectopah.com