Re: Segmentation fault in pg_dumpall from master down to 9.1 and other bug introduced by RLS

On 2014-11-13 18:24:53 -0500, Tom Lane wrote:
> Andres Freund <andres@2ndquadrant.com> writes:
> > On 2014-11-13 18:00:45 -0500, Stephen Frost wrote:
> >> Err.  That might have come off poorly- I didn't mean that sarcastically
> >> but was really wondering how far back you test (or how far back we
> >> feel pg_dumpall needs to work against..).
> 
> > pg_dump currently errors out before 7.0:
> > I do think it might be justifyable to bump that a bit.
> 
> Yeah, we've discussed that before, and I wouldn't mind explicitly ripping
> out support for pre-7.3 or pre-7.4 servers; we could get rid of a pretty
> fair amount of cruft in pg_dump with such a policy change.  But right now
> the standard is 7.0 and I don't want that goalpost moving silently.

Oh, absolutely agreed. It'd be a master only change anyway.

> (And yeah, I did just test it back to 7.0.  But that depends on my HPPA
> box which is looking increasingly frail.  I'm not sure that we can get
> versions that old to compile on more modern platforms, at least not
> without a lot more work than it's probably worth.)

Right. I just tested, and the first one that compiles out of the box on
my quite recent debian sid workstation is 7.4. That also passes make
check.

7.3 fails in compiling the bison output; 7.2 doesn't detect flex; 7.1,
7.0 fail in configure.

Greetings,

Andres Freund

-- Andres Freund                       http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training &
Services