On Fri, Apr 21, 2023 at 12:00:00PM +0300, Alexander Lakhin wrote:
> Please look at the following two bunches for v14+ and v13+ (split to ease
> back-patching if needed). Having processed them, I've reached the state that
> could be considered "clean" ([2], [3]); at least I don't see how to detect
> yet more errors of this class in dozens, so it's my last run for now (though I
> have several entities left, which I couldn't find replacements for).
This was hanging around, and I had some time, so I have looked at the
whole. One of the only two user-visible change was in the docs for
pg_amcheck, so I have applied that first as of 6fd8ae6 and backpatched
it down to 14.
Now, for the remaining 59..
> 1. agg_init_trans_check -> agg_trans
> 2. agg_strict_trans_check -> agg_trans
/*
* pergroup = &aggstate->all_pergroups
- * [op->d.agg_strict_trans_check.setoff]
- * [op->d.agg_init_trans_check.transno];
+ * [op->d.agg_trans.setoff]
+ * [op->d.agg_trans.transno];
*/
Honestly, while incorrect, I have no idea what this comment means ;)
> 4. CommitTSBuffer -> CommitTsBuffer // the inconsistency exists since 5da14938f; maybe this change should be
backpatched
Yes, we'd better backpatch that. I agree that it seems more sensible
here to switch the compiled value rather than what the docs have been
using for years. Perhaps somebody has a different opinion?
The others were OK and in line with the discussion of upthread, so
applied.
--
Michael