Re: set search_path in dump output considered harmful - Mailing list pgsql-hackers

From Marko Kreen
Subject Re: set search_path in dump output considered harmful
Date
Msg-id e51f66da0607180556j1c106b93u840f86f94c1ffec3@mail.gmail.com
Whole thread Raw
In response to Re: set search_path in dump output considered harmful  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
On 7/14/06, Tom Lane <tgl@sss.pgh.pa.us> wrote:

[ problems with missing schema in dump ]

> No, not one of these things can be blamed on pg_dump.

Ok, its not exactly bug but still a big annoyance that
instead dumping fully qualified names it juggles with
search path.  And I'm annoyed as a user looking at the dump:

* To see what schema the function/table/type is in, I need to
carefully look several pages back until there is schema change.

* I cannot copy-paste from dump.

* The function contents must still contain fully-qualified
names, so it cannot be said that the schema is not part of
function definition in some way.  Same goes for other objects.

* The name 'search_path' for picking storage is sick :)
I understand the backwards-compatibility angle, but the over-usage
in dumps makes it even more sick...

-- 
marko


pgsql-hackers by date:

Previous
From: Dave Cramer
Date:
Subject: Re: plPHP and plRuby
Next
From: Joshua Reich
Date:
Subject: Re: [PATCHES] Proposed patch for contrib/cube