Improve error cursor positions for problems with partition bounds.
authorTom Lane <tgl@sss.pgh.pa.us>
Wed, 23 Sep 2020 22:04:53 +0000 (18:04 -0400)
committerTom Lane <tgl@sss.pgh.pa.us>
Wed, 23 Sep 2020 22:04:53 +0000 (18:04 -0400)
commit6b2c4e59d016ea40c42fdc66367d6463e792c125
tree9cadeea7569aed429782e3ec7b4d1dcdf0d31b20
parent3ea7e9550e58c4016625c7e3e4a1340edfc9adf4
Improve error cursor positions for problems with partition bounds.

We failed to pass down the query string to check_new_partition_bound,
so that its attempts to provide error cursor positions were for naught;
one must have the query string to get parser_errposition to do anything.
Adjust its API to require a ParseState to be passed down.

Also, improve the logic inside check_new_partition_bound so that the
cursor points at the partition bound for the specific column causing
the issue, when one can be identified.

That part is also for naught if we can't determine the query position of
the column with the problem.  Improve transformPartitionBoundValue so
that it makes sure that const-simplified partition expressions will be
properly labeled with positions.  In passing, skip calling evaluate_expr
if the value is already a Const, which is surely the most common case.

Alexandra Wang, Ashwin Agrawal, Amit Langote; reviewed by Ashutosh Bapat

Discussion: https://postgr.es/m/CACiyaSopZoqssfMzgHk6fAkp01cL6vnqBdmTw2C5_KJaFR_aMg@mail.gmail.com
Discussion: https://postgr.es/m/CAJV4CdrZ5mKuaEsRSbLf2URQ3h6iMtKD=hik8MaF5WwdmC9uZw@mail.gmail.com
src/backend/commands/tablecmds.c
src/backend/parser/parse_utilcmd.c
src/backend/partitioning/partbounds.c
src/include/partitioning/partbounds.h
src/test/regress/expected/alter_table.out
src/test/regress/expected/create_table.out