[23:35] <wgrant> cjohnston: Do you want me to have a look at the queries in your publisher branch?
[23:35] <wgrant> Bah
[23:35] <wgrant> cjwatson: ^^
[23:36] <cjwatson> wgrant: Oh, if you can, yes please.  I'm not going to have time before tomorrow morning, and possibly not tomorrow in general as I'm on landing team duty.
[23:37] <wgrant> doing
[23:37] <cjwatson> wgrant: I did wonder whether it would be better to just leave the release-pocket-updates checks out of the query and just exclude those after the fact
[23:37] <cjwatson> Since hopefully they're fairly uncommon
[23:38] <wgrant> cjwatson: Not just fairly uncommon, but probably something that should be noticed.
[23:39] <cjwatson> (Do you use EXPLAIN ANALYSE for this kind of thing?)
[23:39] <wgrant> Z, but yes
[23:39] <cjwatson> ANALYZE, sorry.  Bah, Americans
[23:39] <wgrant> Well, EXPLAIN (ANALYZE ON, BUFFERS ON) nowadays.
[23:40] <wgrant> BUFFERS tells you exactly how many pages are read, so you can see if it's just working because of hot caches etc.
[23:40] <cjwatson> Oh, nice