On Sun, Jul 07, 2019 at 08:03:01AM +0300, Alexander Lakhin wrote:
> 5.8. dictlexize -> thesaurus_lexize
There could be other dictionaries.
> 5.9. regression.diffsregression.planregress/inh -> regression.diffs
> planregress/diffs.inh
I am wondering if we should not just nuke that... For now I have
included your change as the mistake is obvious, but I am starting a
new thread. The history around this script does not play in favor of
it:
commit: 2fc80e8e8304913c8dd1090bb2976632c0f4a8c3
author: Bruce Momjian <bruce@momjian.us>
date: Wed, 12 Feb 2014 17:29:19 -0500
Rename 'gmake' to 'make' in docs and recommended commands
This simplifies the docs and makes it easier to cut/paste command
lines.
commit: c77e2e42fb4cf5c90a7562b9df289165ff164df1
author: Tom Lane <tgl@sss.pgh.pa.us>
date: Mon, 18 Dec 2000 02:45:47 +0000
Tweak regressplans.sh to use any already-set PGOPTIONS.
And looking closer it seems that there are other issues linked to
it...
> 5.27. equivalentOpersAfterPromotion -> remove (irrelevant since
> 8536c962, but the whole comments is too old to be informational too)
It seems to me that this could be much more reworked. So discarded
for now.
> 5.29. ExclusiveRowLock -> RowExclusiveLock
Grammar mistake here.
> 5.31. ExecBitmapHeapNext -> BitmapHeapNext
Well, BitmapHeapRecheck is not listed in the interface routines
either..
> 5.37. ExecSeqNext -> SeqNext
> 5.40. ExecSubqueryNext -> SubqueryNext
> 5.41. ExecValuesNext -> ValuesNext
Here as well these sets are incomplete. Instead for those series I'd
like to think that it would be better to do a larger cleanup and just
remove all these in the executor "INTERFACE ROUTINES". Your proposed
patches don't make things better either, as the interfaces are listed
in alphabetical order.
> 5.39. exec_subplan_get_plan -> remove (not used since 1cc29fe7)
This could be used by extensions. So let's not remove it.
And committed most of the rest. Thanks.
--
Michael