> > pg_dump gives me back :-
> > CREATE TYPE sotime ( internallength = 2, externallength = 5,
> > input = sotime_in_18272, output = sotime_out_18304, send = sotime_out_18304,
> > receive = sotime_in_18272, default = '-',passedbyvalue);
> > CREATE TYPE _sotime ( internallength = -1, externallength = -1,
> > input = array_in_750, output = array_out_751, send = array_out_751,
> > receive = array_in_750, default = '-');
> Yes, we have a problem with user-defined types because of the changes
> I made to make the function name unique by adding the pg_proc oid to
> the name.
> I can make the fix, but no one seems to have an opinion on how yet.
It would seem that for Keith's cases things worked OK originally? There
was trouble with other overloaded functions?
At worst, we should revert to the non-mangled names. On a related note,
has anyone had time to try testing pg_upgrade? That would be a nice
feature for v6.4 if it worked on a wide range of databases (haven't
tried it myself yet).
In response to
pgsql-hackers by date
|Next:||From: Thomas G. Lockhart||Date: 1998-09-24 06:03:56|
|Subject: Re: [HACKERS] pg_dump, problem with user defined types?|
|Previous:||From: Thomas G. Lockhart||Date: 1998-09-24 02:34:24|
|Subject: Re: [HACKERS] Re: [SQL] 2 questions.|