Gavin Sherry wrote:
> Here's one of the patches I was refering to
>
> ---------- Forwarded message ----------
> Date: Mon, 15 May 2006 23:22:33 +1000 (EST)
> From: Gavin Sherry <swm at alcove.com.au>
> To: slony1-general at gborg.postgresql.org
> Subject: Slony with PostgreSQL mainline
>
> Hi guys,
>
> Tom modified the signature of typenameTypeId() so as to implement his
> error offset handling when type lookups fail. The signature is now:
>
> typenameTypeId(ParseState *pstate, const TypeName *typename)
>
> For our purposes, we can just pass NULL for the first argument. In
> mainline PostgreSQL this is fine because pstate is not dereferenced
> without first testing if it is equal to NULL. Also, pstate is only used if
> the type can not be found. We've got other problems if that's not possible
> at the two call sites inside the slony code.
>
> An attached patch updates the code to handle the API change.
>
>   
An equivalent change went in already...  It's possible that yours is a
bit nicer, but I think this already did the trick...

<http://gborg.postgresql.org/pipermail/slony1-commit/2006-June/001045.html>