Re: Revised signal multiplexer patch - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Revised signal multiplexer patch
Date
Msg-id 24211.1249062023@sss.pgh.pa.us
Whole thread Raw
In response to Revised signal multiplexer patch  (Fujii Masao <masao.fujii@gmail.com>)
List pgsql-hackers
Fujii Masao <masao.fujii@gmail.com> writes:
> But there is one issue; the extra search is always required to send a notify
> interrupt. This is because pg_listener doesn't have a backend ID and we
> cannot pass it to SendProcSignal. In order to solve this issue, we should
> newly add backend ID field into pg_listener?

Hmm.  I'm not tremendously concerned about that --- the LISTEN/NOTIFY
code has been on the agenda for a complete rewrite for a long time now,
and I keep hoping pg_listener will go away entirely sometime soon.
I don't feel a need to go and fix a marginal performance issue there.
        regards, tom lane


pgsql-hackers by date:

Previous
From: "Kevin Grittner"
Date:
Subject: Re: Review: Revise parallel pg_restore's scheduling heuristic
Next
From: Magnus Hagander
Date:
Subject: Re: 8.4 win32 shared memory patch