On Sep 15, 2011, at 10:09 AM, Ross Peoples wrote: > I think we can rule out the json serializer. I just replaced the code in > tools.py and serializers.py and tried this. I also confirmed that the C json > library was installed. There wasn't any change. However, I wonder if I should > send those changes to Massimo as a patch anyways since it seems like we > should prefer the C libraries when possible.
I think that'd be a good idea. However, please covert it to using the name 'json' instead of 'simplejson', if you haven't already done that. This is a good time to do it, because it's getting some good testing. > > As for your other question, simply having "auth = Auth(db)" in there without > defining the tables seems to slow down XML by a few ms, but we are talking > 5ms at most. The define_tables call is what slows it down the other 20 ms. That makes sense. There's one table definition in the Auth(db) call. > > However, it may not be Auth at all. I just defined a dummy table just to see > what would happen, and the XML times went back up to 50 ms, so it looks like > the DAL is somehow causing the slowdown. And I just run a second test on the > dummy table with "migrate=False" and the numbers dropped back down to 20ms. > > So it looks like we've narrowed down the problem to migrations. And through all this the JSON timings stayed high?