> Yes, this is exactly the concern I raised last week. Thomas didn't
> seem to be very worried about the issue, but when he gets back from
> his vacation we can lean on him to fix it.
OK Tom I'll try to sound more concerned next time :))
I'm using the rte->ref Attr structure to carry internal info on table
names and column names. What I should be able to do is decouple the
internal ref structure from the table name/column list specified by a
user, so the "query recreation" code can ignore the internal structure
and just use the original list from the user.
Should be able to go into v7.0 with no problem (other than initdb, but
it *is* a beta!!).
> Something else we might consider as a stopgap is to resurrect the
> "compressed text" datatype that Jan wrote, and then removed in
> anticipation of having TOAST. Jan was concerned about creating
> future compatibility problems by having a datatype with only a
> one-release-cycle expected lifetime ... but I think it might be
> OK to use it just internally for rules.
Naw, the above should be easier all around.
- Thomas
--
Thomas Lockhart lockhart@alumni.caltech.edu
South Pasadena, California
