Thread: pgsql: In opr_sanity regression test, check for unexpected uses of cstr

pgsql: In opr_sanity regression test, check for unexpected uses of cstr

From
Tom Lane
Date:
In opr_sanity regression test, check for unexpected uses of cstring.

In light of commit ea0d494dae0d3d6f, it seems like a good idea to add
a regression test that will complain about random functions taking or
returning cstring.  Only I/O support functions and encoding conversion
functions should be declared that way.

While at it, add some checks that encoding conversion functions are
declared properly.  Since pg_conversion isn't populated manually,
it's not quite as necessary to check its contents as it is for catalogs
like pg_proc; but one thing we definitely have not tested in the past
is whether the identified conproc for a conversion actually does that
conversion vs. some other one.

Branch
------
master

Details
-------
http://git.postgresql.org/pg/commitdiff/921191912c48a68db81c02c02f3bc22e291d918c

Modified Files
--------------
src/test/regress/expected/opr_sanity.out |   84 +++++++++++++++++++++++++++++-
src/test/regress/sql/opr_sanity.sql      |   73 +++++++++++++++++++++++++-
2 files changed, 155 insertions(+), 2 deletions(-)