]> granicus.if.org Git - postgresql/commit
Fix several mistakes around parallel workers and client_encoding.
authorRobert Haas <rhaas@postgresql.org>
Thu, 30 Jun 2016 22:35:32 +0000 (18:35 -0400)
committerRobert Haas <rhaas@postgresql.org>
Thu, 30 Jun 2016 22:35:32 +0000 (18:35 -0400)
commit10c0558ffefcd12bf1d3dc35587eba41d1ce4571
tree5b6344c420e2822f7c66f626294218adbd54343f
parentf8c58554db48fe004938a8a34a42afb78157b66c
Fix several mistakes around parallel workers and client_encoding.

Previously, workers sent data to the leader using the client encoding.
That mostly worked, but the leader the converted the data back to the
server encoding.  Since not all encoding conversions are reversible,
that could provoke failures.  Fix by using the database encoding for
all communication between worker and leader.

Also, while temporary changes to GUC settings, as from the SET clause
of a function, are in general OK for parallel query, changing
client_encoding this way inside of a parallel worker is not OK.
Previously, that would have confused the leader; with these changes,
it would not confuse the leader, but it wouldn't do anything either.
So refuse such changes in parallel workers.

Also, the previous code naively assumed that when it received a
NotifyResonse from the worker, it could pass that directly back to the
user.  But now that worker-to-leader communication always uses the
database encoding, that's clearly no longer correct - though,
actually, the old way was always broken for V2 clients.  So
disassemble and reconstitute the message instead.

Issues reported by Peter Eisentraut.  Patch by me, reviewed by
Peter Eisentraut.
src/backend/access/transam/parallel.c
src/backend/commands/async.c
src/backend/commands/variable.c
src/backend/libpq/pqformat.c
src/backend/libpq/pqmq.c
src/include/commands/async.h
src/include/libpq/pqformat.h