Tom,
* Tom Lane (tgl@sss.pgh.pa.us) wrote:
> Given that this has been like this right along, I don't see why it's
> all that urgent to force a half-baked solution into 9.1. I'm also
> concerned that if we do do that, you'll lose motivation to work on
> cleaning it up for 9.2 ;-)
The addition to log_line_prefix is hardly 'half-baked' as a solution, to
that problem (I just pulled the hunks from the rest of the patch as they
were completely independent, and tested them). I've also gone and added
the csvlog_fields/csvlog_header patch to the 2011-Next commitfest. :P
I've also already started looking at changing syslogger to have it
figure out if it should be writing a header out or not. If we can
decide what semantics we should have when the log file exists and we're
not planning to rotate it on startup, it won't be hard for me to
implement them (well, I hope).
Thanks,
Stephen