Re: pg_dump gets attributes from tables in extensions

On Tue, Feb 24, 2015 at 11:55 PM, Alvaro Herrera <alvherre(at)2ndquadrant(dot)com>
wrote:

> Michael Paquier wrote:
> > On Tue, Feb 24, 2015 at 3:13 PM, Rushabh Lathia <
> rushabh(dot)lathia(at)gmail(dot)com>
> > wrote:
>
> > > rushabh(at)rushabh-centos-vm:dump_test$ cat dump_test--1.0.sql
> > > /* dump_test/dump_test--1.0.sql */
>
> Hm. I think it would be a good idea to collect these extension files
> somewhere so that pg_dump hacking can be tested with them. Right now,
> the extension part of pg_dump is easy to break in subtle ways, which
> discourages anyone who wants to meddle with it.
>
> Maybe something in src/test/modules could keep these files so that
> pg_dump can be tested. Is anybody interested in doing that?
>

For the patch to fix data dump of extensions that contain tables with FK I
have implemented a test case in src/test/modules as I didn't want to look
at postgis stuff to reproduce the failure, last version is here:
http://www.postgresql.org/message-id/CAB7nPqQRjHHmNMcGB0eCVZ2PTMNCmFD+1htLOoxRQjMzgtSetQ@mail.gmail.com

> Btw, perhaps you may have noticed, but I marked this patch as rejected...
> I
> > don't think it makes much sense to put restrictions in this code path
> after
> > finding my way through all the stuff of pg_dump.
>
> Please don't do that -- I mean don't use the commitfest app as the only
> source for random bits of data. If you want to reject a patch, please
> make sure to point that out in the email thread. It's impossible to go
> from email archives to commitfest entries.
>

Sure. Sorry for that...
--
Michael