Bruce Momjian <bruce(at)momjian(dot)us> writes:
> Tom Lane wrote:
>> One point after looking back at the previous discussion is that the
>> current version test is too strict: it will complain if your server is
>> 8.2.7 and pg_dump is 8.2.6. We probably should not make a newer minor
>> number a hard error, since 99.99% of the time it would be fine. So
>> while I think newer major should be a hard error regardless of -i,
>> we could consider several responses to newer minor:
>> * silently allow it always
>> * print warning and proceed always
>> * allow -i to control error vs warning for this case only.
> I think it should be silent. Do we ever change the server behavior that
> is visible to pg_dump in a minor release?
It's hardly out of the question --- consider the backslash-escaping
security fixes we applied in 8.1.4, 8.0.8, etc. Parts of the server
changes were intended to intentionally break unpatched clients, and
I think that'd apply to unpatched pg_dump as well.
Of course, that precedent suggests that any such change would be made in
such a way as to be enforced on the server side, so it wouldn't matter
if pg_dump didn't know it wouldn't work.
Silent allow is fine with me, I was just wondering if anyone liked
the other options better.
regards, tom lane
In response to
pgsql-hackers by date
|Next:||From: Gurjeet Singh||Date: 2008-03-26 19:01:29|
|Subject: Re: HELP|
|Previous:||From: Bruce Momjian||Date: 2008-03-26 18:46:30|
|Subject: Re: Re: [COMMITTERS] pgsql: Strengthen warnings
about using pg_dump's -i option.|
pgsql-committers by date
|Next:||From: Alvaro Herrera||Date: 2008-03-26 21:10:39|
|Subject: pgsql: Move the HTSU_Result enum definition into snapshot.h, to avoid |
|Previous:||From: Alvaro Herrera||Date: 2008-03-26 18:49:00|
|Subject: pgsql: Rename snapmgmt.c/h to snapmgr.c/h, for consistency with other |