Re-stabilize infinite_recurse() test case.
Since commit 8f59f6b9c0, CLOBBER_CACHE_ALWAYS buildfarm members have
been failing this test case because the error message now sometimes
includes an error cursor position. It seems largely just luck that
that never happened before, and there are likely to be more ways it
could happen in future. Hence, rather than trying to prevent it,
adjust the test script to suppress that component of the report.
At some point we might need to back-patch this, but refrain until
there's a demonstrated need. (We'd need a different fix before v12,
anyway, since VERBOSITY=sqlstate is a recent thing.)
Tom Lane and Andres Freund
Discussion: https://postgr.es/m/30675.1586111599@sss.pgh.pa.us
Branch
------
master
Details
-------
https://git.postgresql.org/pg/commitdiff/3c8553547b1493c4afdb80393f4a47dbfa019a79
Modified Files
--------------
src/test/regress/expected/errors.out | 9 ++++++---
src/test/regress/sql/errors.sql | 6 ++++--
2 files changed, 10 insertions(+), 5 deletions(-)
