Re: [HACKERS] Remove add_missing_from_clause?

"Joshua D. Drake" <jd(at)commandprompt(dot)com> writes:
> Tom Lane wrote:
>> No, AFAICT it's still in active use --- you need not look back far in
>> the mailing lists to find questions answered by "see add_missing_from".

> Oh, I know it is still in use, that is my complaint :). Perhaps stating
> that this will be the last release of the feature?

Stating it doesn't make it so ;-)

I actually don't see a big problem with add_missing_from. It's not a
security hole because (a) it doesn't affect properly written queries and
(b) the worst you can do to an improperly written query is make it fail
in an obvious way. So I'm willing to leave it around for legacy apps
for a good while; certainly as long as we left sql_inheritance in there,
which will have been 6+ years if it goes away in 8.3. The clock on
add_missing_from can only be said to have started running when we made
it false by default, which was 8.1 --- so it's got about five years left...

Also, I see in the 8.1 release notes

It may be necessary to set add_missing_from to true in order to
load an existing dump file, if the dump contains any views or rules
created using the implicit-FROM syntax. This should be a one-time
annoyance, because PostgreSQL 8.1 will convert such views and rules
to standard explicit-FROM syntax. Subsequent dumps will therefore
not have the problem.

If we remove add_missing_from then some people will be unable to migrate
forward from pre-8.1 releases because their views won't load. Seeing
that you have elsewhere been arguing for the wisdom of skipping
releases, I can hardly think that you'd be in favor of that.

regards, tom lane