Обсуждение: Re: [GENERAL] Using Cursor in PostgreSQL 7.2

Tom Lane wrote:

> I've added the missing checks in the planner; possibly you could get rid
> of that UNION hack now.
> 


*Moved to hackers*

I confirmed the UNION hack is no longer required. Thanks! Is it too late 
to change the README in contrib/dblink?


A side issue I noticed is that recent changes to contrib/*/Makefile seem 
to cause 'MODULE_PATHNAME' in *.sql.in files to become 
'$libdir/modulename' in the resulting *.sql files. Example:

in rtree_gist.sql.in:
-- define the GiST support methods
create function gbox_consistent(opaque,box,int4) returns bool as 
'MODULE_PATHNAME' language 'C';

becomes in rtree_gist.sql:
-- define the GiST support methods
create function gbox_consistent(opaque,box,int4) returns bool as 
'$libdir/rtree_gist' language 'C';

Same thing happens in (at least) dblink.sql, fuzzystrmatch.sql, and 
array_iterator.sql.

I'm not sure right off how to fix it though :(

Joe

Joe Conway <joseph.conway@home.com> writes:
> I confirmed the UNION hack is no longer required. Thanks! Is it too late 
> to change the README in contrib/dblink?

No, I don't think that's a problem.  Send a patch.

> A side issue I noticed is that recent changes to contrib/*/Makefile seem 
> to cause 'MODULE_PATHNAME' in *.sql.in files to become 
> '$libdir/modulename' in the resulting *.sql files. Example:

This is correct behavior now; in fact, full paths in CREATE FUNCTION
should be deprecated...
        regards, tom lane