On 2018-11-26 15:09:43 -0300, Alvaro Herrera wrote: > On 2018-Nov-26, Stephen Frost wrote: > > > I would think we'd have the different GUCs and then the check functions > > would only validate that they're valid inputs and then when we get to > > the point where we're starting to do recovery we check and make sure > > we've been given a sane overall configuration- which means that only > > *one* is set, and it matches the recovery target requested. > > I don't quite understand why it isn't sensible to specify more than one > and just stop recovery (or whatever) when at least one of them becomes > true. Maybe I want to terminate just before commit of transaction > 12345, but no later than 2018-11-11 12:47 in any case.
I really have doubts about that being a serious use-case. If you know the xid then you almost certainly want everything before that xid, and you use the time stamp when you don’t know the xid (which is pretty much always..).
+1
Well, we could start with: that isn’t how things work today, nor how it used to work before this patch, and we’ve not had anyone asking for it except for people on this thread making things up.
Let’s at least fix the currently committed patch before adding new features or changing how things in recovery work.