> >> Hrm, alright, well, with the coming of 7.3.3 soon-ish, and given
> >> that RT and Horde are still broken (does anyone have any pull
> >> with that crowd and the ability to submit patches?), I'll make
> >> sure to conditionalize patch when the next version comes out, but
> >> I'm not going to force admins to rebuild their DBs unless really
> >> necessary.
>
> > RT's author hasn't released a release in a LONG time, and I don't
> > have any pull.
>
> > As to Horde, I don't have any insight. Sorry.
>
> The obvious solution to this sort of problem would be to introduce a
> GUC variable ("allow_zero_length_integers" or some such name). I do
> not recall why we decided not to do that to begin with --- should we
> reconsider, or is it too late to worry about it?
I'd bury it behind a #define + configure option, if anything. I
wouldn't want any runtime cycles spent on something like this.
I hate to take the side of legacy and know it'd be feature thrash, but
I'd think it'd be appropriate to have this be a WARNING for the life
of 7.3.X and then make it an error in 7.4 where major changes/breakage
like this are expected. -sc
--
Sean Chittenden