Bruce Momjian wrote:
> This is normally caused by Stephan's patches. His patches were
> originally against 6.3, and he ported them to 6.4, but he normally does
> lots of development without any communication with us, sends us a huge
> patch, and we normally have to clean up the edges somewhat. This patch
> actually caused fewer problems than the HAVING patch he submitted.
Porting his patch to v6.4 was not exactly what he did. He
changed the v6.5 tree in a way that his patch fit's into.
Namely he removed copyObject() in some cases. copyObject() is
an expensive function. There are only 2 reasons to call it.
One is that the object in question lives in a memory context
that could get destroyed before we are done with the object.
The other is that we are about to change the object but
others need it unchanged (in the actual case varno's got
changed).
And he reverted the possibility to group multiple rule
actions in ()'s.
One good thing it caused is, that I realize I was wrong!
LIMIT seems to never have been applied to the tree - OOOPS. I
don't know how this could have happened. Must do it before
v6.5 BETA because it's FEATURE.
Jan
--
#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me. #
#======================================== jwieck@debis.com (Jan Wieck) #