William Garrison <postgres@mobydisk.com> writes:
> Tom Lane wrote:
>> What do you mean by "one batch" exactly? Both CREATE and DROP TABLESPACE
>> refuse to run in a transaction block, so I'm confused about this.
> Not a transaction block. A batch of commands submitted to the server in
> a single call. In MSSQL land, I call that a batch. I don't know the
> PostgreSql term. If you are using the pgadmin3 GUI, then I mean pressing
> F5 once is a single batch. Pressing it twice is two batches.
Oh, it's a single simple-Query message. You could reproduce the problem
with psql if you put multiple commands into a "-c" command line switch.
This is a basic oversight in PreventTransactionChain: it doesn't reject
the case where the command is submitted as part of a multi-query string
in a single Query message.
This is relatively easy to fix in CVS HEAD --- we can just teach
exec_simple_query to pass isTopLevel = true only when the querystring
contains a single command, or maybe better only for the last command
of a querystring. I don't see any very practical way to fix it in
older releases though; at least not anything I'd want to backpatch
when it can't be tested first in HEAD. Anyone have an idea about a
reasonable back-branch fix?
regards, tom lane