> * You were using a bare C string as the amname argument in COMMENT ON
> OPERATOR CLASS. This won't do because the parse tree is not a valid
> Node structure; copyObject will fail on it. I inserted makeString()
> and strVal() calls to fix it.
>
> BTW, a simple test to detect uncopiable-parsetree problems is to compile
> with COPY_PARSE_PLAN_TREES defined. Doing so revealed that you're not
> the only person to have made this mistake lately --- ALTER SEQUENCE is
> broken too.
Ok.
> * I made the macros LARGE and OBJECT be LARGE_P and OBJECT_P; they
> seemed just a little too ripe for conflicts as-is ...
>
> * The pg_dump code for COMMENT ON OPCLASS pretty obviously had not been
> tested :-(
Hrm. Yeah, weird. I think I just forgot, because I learned in this
patch that you can't do two of those %s in the append function, and I
had fixed it in all the other dumps. *sigh*
The other thing I was concerned about was a bit of code duplication,
especially for the comment on opclass function.
Out of interest, I notice you didn't commit my inv_api.c change to
delete comments on LOBs - where did you put it instead?
Chris