Re: [HACKERS] config.h/Followup FOLLOWUP - Mailing list pgsql-hackers

From darrenk@insightdist.com (Darren King)
Subject Re: [HACKERS] config.h/Followup FOLLOWUP
Date
Msg-id 0b456a25b820668a99a3955245343a53
Whole thread Raw
List pgsql-hackers
>
> So, is it lp_len or lp_offset that can be reduced by 2?  I want to
> experiment with this...
>

Neither...try grep'ing around for uses of lp_flags.  I dug into this
last Dec/Jan...check the hackers digests from that time for any
relevent info.  At that time, only two bits in lp_flags were in use.
Don't know if any more are taken now or not.

Both lp_len and lp_offset should be the same, so if you take four bits
from lp_flags (and give two apiece to lp_len & lp_offset), that would
get you to a block size of 32k.

Now that there're timely src snapshots available, I'm going to try to
get back into coding (assuming the aix port still works. :)

darrenk

------------------------------

pgsql-hackers by date:

Previous
From: "Sergey V. Lunegov"
Date:
Subject: [HACKERS] subsribe
Next
From: "Vadim B. Mikheev"
Date:
Subject: Re: [HACKERS] subselect