© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
f33e652
(Get the code!)
Right. *my* point is that the xmlrpc being run on "locally" is issuing 3 total queries, but takes up more than 700ms on my VM. No good way to test that on production, I think. I suppose I could add some debug logging that reports how long the XMLRPC calls are taking as part of lp-serve.
Given that the *bzr* work to actually initialize the directory was only 200ms, spending 3x longer to decide that it could actually do them seems significant, and worth trying to optimize.
As Robert says, I'm sure the actual datacenter XMLRPC logic in production is horribly complex and might be hard to optimize. I might give it a poke and see if there is anything obvious. (SQL request on a table without the appropriate index, etc.)