Re: [HACKERS] Possible explanation for Win32 stats - Mailing list pgsql-patches

From korryd@enterprisedb.com
Subject Re: [HACKERS] Possible explanation for Win32 stats
Date
Msg-id 1154201708.27274.17.camel@sakai.localdomain
Whole thread Raw
In response to Re: [HACKERS] Possible explanation for Win32 stats regression  (Andrew Dunstan <andrew@dunslane.net>)
Responses Re: [HACKERS] Possible explanation for Win32 stats regression  (Andrew Dunstan <andrew@dunslane.net>)
List pgsql-patches
heh. I was just doing it the way Tom suggested - see attached. With a 
little more trouble we could also keep track if the listened for events 
and sometimes save ourselves a second call to WSAEventSelect, but I'm 
not sure it's worth it.

It all depends on the overhead of WSAEventSelect().  I'm sure your version would run faster, but I just don't know if "slower" would be measurable.

BTW:  I would suggest changing your comment to:

       /*
        * make sure we don't multiplex this kernel event object with a different socket
        * from a previous call
        */

Thanks for tackling this problem too.

    -- Korry


----
  Korry Douglas    korryd@enterprisedb.com
  EnterpriseDB      http://www.enterprisedb.com

pgsql-patches by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: CREATE INDEX ... ONLINE
Next
From: Andrew Dunstan
Date:
Subject: Re: [HACKERS] Possible explanation for Win32 stats regression