On Friday, May 6, 2016, Tom Lane <t...@sss.pgh.pa.us> wrote: > If you're not tired of reviewing release notes (I'm sure getting a bit > tired of writing them), see > > > http://git.postgresql.org/gitweb/?p=postgresql.git;a=commitdiff;h=eb7de00ac2d282263541ece849ec71e2809e9467 > > guaibasaurus should have 'em up on the web in an hour or so, too, at > > http://www.postgresql.org/docs/devel/static/release-9-5-3.html > > "...replacement_sort_tuples, which see for further details." needs rewording.
For some reason I had trouble comprehending the index only scans on partial index couple or paragraphs. Got it after a few reads. Seems like it's almost too detailed. "Partial indexes can be used for index only scans in some circumstances. See section for details." If there isn't a section to point to there should be - people want to know how to get IOS and aren't going to read release notes to figure it out. Are the pg_stat_activity changes breaking changes? If so its not clear from the notes. I'll +1 the elsewhere mentioned confusion adding a pg_config view vis-a-vis pg_settings. Adding (or using) the word "compile" would be advisable. The guc for the number of standby servers that must acknowledge should be named in the notes and linked to the main docs. "An additional syntax has been added to synchronous_standby_names to accommodate the number of standby servers that must acknowledge a commit." Is it worth mentioning the deprecation of exclusive backups in the notes introducing non-exclusive ones? Read the rest and nothing stood out - though I guess I'd advise myself or the next person to read up from the bottom so fresh eyes read the lower stuff first. David J.