]> granicus.if.org Git - postgresql/commit
Give up on testing guc.c's behavior for "infinity" inputs.
authorTom Lane <tgl@sss.pgh.pa.us>
Mon, 11 Mar 2019 21:53:09 +0000 (17:53 -0400)
committerTom Lane <tgl@sss.pgh.pa.us>
Mon, 11 Mar 2019 21:53:09 +0000 (17:53 -0400)
commitd9c5e9629bf511a51328fd083ee452de88d91d9d
tree36d511b5e903f6e5410b8d06b5c4cfdddabc188b
parent8cacea7a725103f1a037a5ee06112ebe31051c66
Give up on testing guc.c's behavior for "infinity" inputs.

Further buildfarm testing shows that on the machines that are failing
ac75959cd's test case, what we're actually getting from strtod("-infinity")
is a syntax error (endptr == value) not ERANGE at all.  This test case
is not worth carrying two sets of expected output for, so just remove it,
and revert commit b212245f9's misguided attempt to work around the platform
dependency.

Discussion: https://postgr.es/m/E1h33xk-0001Og-Gs@gemulon.postgresql.org
src/backend/utils/misc/guc.c
src/test/regress/expected/guc.out
src/test/regress/sql/guc.sql