> -----Original Message----- > From: Darren Shepherd [mailto:dar...@godaddy.com] > Sent: Monday, August 13, 2012 1:29 PM > To: cloudstack-dev@incubator.apache.org > Subject: RE: Hibernate -> Custom DAO for AWS component > > I think there's two points. > > 1) Raw SQL is bad, so even in a time crunch moving to the custom CloudStack > DAO implementation *should* have been better. > 2) I'd rather see a JPA based solution then moving to custom CloudStack DAO. > > I think Alex and I agree on point 1, and I'm not sure if anybody agrees with > point 2. What's the time frame for when this was supposed to be done?
I agree on 1. I don't really have much of opinion on two other than the current DAO code works pretty well so I don't think it's a priority to swap it out. But I do want to point out the difference between cloudstack's orchestration code and api code. I think the api code should not have used the custom DAO. The api does much more joining and acl control. So to me it makes more sense that the api code switch to a better search framework than the custom DAO. If anything we should start there to switch out the DAO layer. --Alex
