]> granicus.if.org Git - postgresql/commit
Make SPI_fnumber() reject dropped columns.
authorTom Lane <tgl@sss.pgh.pa.us>
Tue, 8 Nov 2016 18:11:15 +0000 (13:11 -0500)
committerTom Lane <tgl@sss.pgh.pa.us>
Tue, 8 Nov 2016 18:11:26 +0000 (13:11 -0500)
commit6d30fb1f75a57d80f80e27770d39d88f8aa32d28
tree056042db35fa8ad69957dd54dda3c5dcaaa62d7c
parent36ac6d0e793087153a452df6502d0ef32a780db6
Make SPI_fnumber() reject dropped columns.

There's basically no scenario where it's sensible for this to match
dropped columns, so put a test for dropped-ness into SPI_fnumber()
itself, and excise the test from the small number of callers that
were paying attention to the case.  (Most weren't :-(.)

In passing, normalize tests at call sites: always reject attnum <= 0
if we're disallowing system columns.  Previously there was a mixture
of "< 0" and "<= 0" tests.  This makes no practical difference since
SPI_fnumber() never returns 0, but I'm feeling pedantic today.

Also, in the places that are actually live user-facing code and not
legacy cruft, distinguish "column not found" from "can't handle
system column".

Per discussion with Jim Nasby; thi supersedes his original patch
that just changed the behavior at one call site.

Discussion: <b2de8258-c4c0-1cb8-7b97-e8538e5c975c@BlueTreble.com>
contrib/spi/autoinc.c
contrib/spi/insert_username.c
contrib/spi/moddatetime.c
contrib/spi/refint.c
contrib/spi/timetravel.c
doc/src/sgml/spi.sgml
src/backend/executor/spi.c
src/backend/utils/adt/tsvector_op.c
src/pl/plperl/plperl.c
src/pl/tcl/pltcl.c
src/test/regress/regress.c