Undo decision to allow pg_proc.prosrc to be NULL.
authorTom Lane <tgl@sss.pgh.pa.us>
Thu, 15 Apr 2021 21:17:20 +0000 (17:17 -0400)
committerTom Lane <tgl@sss.pgh.pa.us>
Thu, 15 Apr 2021 21:17:20 +0000 (17:17 -0400)
commit1111b2668d89bfcb6f502789158b1233ab4217a6
tree38e331ef19b32e276c231ad91fc4ae07327d6ceb
parent3157cbe974846729d49a1ee081944eee1839bdd8
Undo decision to allow pg_proc.prosrc to be NULL.

Commit e717a9a18 changed the longstanding rule that prosrc is NOT NULL
because when a SQL-language function is written in SQL-standard style,
we don't currently have anything useful to put there.  This seems a poor
decision though, as it could easily have negative impacts on external
PLs (opening them to crashes they didn't use to have, for instance).
SQL-function-related code can just as easily test "is prosqlbody not
null" as "is prosrc null", so there's no real gain there either.
Hence, revert the NOT NULL marking removal and adjust related logic.

For now, we just put an empty string into prosrc for SQL-standard
functions.  Maybe we'll have a better idea later, although the
history of things like pg_attrdef.adsrc suggests that it's not
easy to maintain a string equivalent of a node tree.

This also adds an assertion that queryDesc->sourceText != NULL
to standard_ExecutorStart.  We'd been silently relying on that
for awhile, so let's make it less silent.

Also fix some overlooked documentation and test cases.

Discussion: https://postgr.es/m/2197698.1617984583@sss.pgh.pa.us
15 files changed:
doc/src/sgml/catalogs.sgml
src/backend/catalog/pg_proc.c
src/backend/commands/functioncmds.c
src/backend/executor/execMain.c
src/backend/executor/functions.c
src/backend/optimizer/util/clauses.c
src/bin/pg_dump/pg_dump.c
src/bin/psql/describe.c
src/include/catalog/catversion.h
src/include/catalog/pg_proc.h
src/include/executor/functions.h
src/test/regress/expected/create_function_3.out
src/test/regress/expected/opr_sanity.out
src/test/regress/sql/create_function_3.sql
src/test/regress/sql/opr_sanity.sql