Re: "interesting" issue with restore from a pg_dump with a database-wide search_path

"Regina Obe" <lr(at)pcorp(dot)us> writes:
> Sorry for not posting from the thread. Paul alerted me to this one and I am aware of the issue.

> 1) I do have fuzzstrmatch listed as a dependency in the control file. I know because I often install the geocoder with
> CREATE EXTENSION postgis_tiger_geocoder CASCADE;
> And it installs postgis and fuzzystrmatch

OK.

> a) In this particular case, I have a function that uses fuzzystrmatch and is used in functional indexes.
> I unfortunately can't schema qualify the use of soundex, because I don't know where the user may have installed fuzzystrmatch is installed
> b) Stephen Frost had suggested, perhaps we should have some syntax like @extension_loc(fuzzystrmatch)...@ so that one could reference an extension dependency location within a function without knowing where it is installed.

You don't really need any new syntax for this particular case, I think.
You can declare the function in the extension like this:

create function ... set search_path from current;

which will cause it to absorb the search path that's set while running
the extension script, which should be what you want.

regards, tom lane