]> granicus.if.org Git - postgresql/commit
Fix canAcceptConnections() bugs introduced by replication-related patches.
authorTom Lane <tgl@sss.pgh.pa.us>
Sun, 14 Nov 2010 20:57:37 +0000 (15:57 -0500)
committerTom Lane <tgl@sss.pgh.pa.us>
Sun, 14 Nov 2010 20:57:37 +0000 (15:57 -0500)
commit3892a2d8619ee5623fc3ebee775538fdea1f529e
treecb82d49335164ca5a9f7126a6808c8e190b2b685
parent1bd201214965b3eba23bee2c4ce4c14accde170f
Fix canAcceptConnections() bugs introduced by replication-related patches.

We must not return any "okay to proceed" result code without having checked
for too many children, else we might fail later on when trying to add the
new child to one of the per-child state arrays.  It's not clear whether
this oversight explains Stefan Kaltenbrunner's recent report, but it could
certainly produce a similar symptom.

Back-patch to 8.4; the logic was not broken before that.
src/backend/postmaster/postmaster.c