Re: Need help understanding pg_locks

Bruce Momjian <bruce@momjian.us>

From: Bruce Momjian <bruce@momjian.us>
To: Florian Pflug <fgp@phlo.org>
Cc: Andrew Dunstan <andrew@dunslane.net>, Tom Lane <tgl@sss.pgh.pa.us>, PostgreSQL-development <pgsql-hackers@postgresql.org>
Date: 2011-07-14T20:18:09Z
Lists: pgsql-hackers

Attachments

Florian Pflug wrote:
> I still believe the chance of confusion to be extremely small, but since
> you feel otherwise, what about "Targeted" instead of "Locked". As in
> 
>   OID of the relation targeted by the lock, or null if the lock does not
>   target a relation or part of a relation.
> 
>   Page number within the relation targeted by the lock, or null if the
>   lock does not target a tuple or a relation page.
> 
>   Virtual ID of the transaction targeted by the lock, or null if the lock
>   does not target a virtual transaction ID.
> 
> "Protected"/"protects" instead of "Targeted"/"targets" would also work.
> 
> Both avoid the imprecision of saying "Locked", and the ambiguity "on" -
> which might either mean the physical location of the lock, or the object
> its protecting/targeting. 
> 
> > I reworded that line to:
> > 
> > +       OID of the relation of the lock target, or null if the lock is not
> 
> I'm not a huge fan of that. IMHO " .. of .. of .. " chains are hard to
> read. Plus, there isn't such a thing as the "relation of a lock target" -
> the relation *is* the lock target, not a part thereof.

Agreed.  I like "targeted by".  New patch attached.

-- 
  Bruce Momjian  <bruce@momjian.us>        http://momjian.us
  EnterpriseDB                             http://enterprisedb.com

  + It's impossible for everything to be true. +