> On Oct 12, 2021, at 10:03 AM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
>
> On Tue, Oct 12, 2021 at 12:57 PM Justin Pryzby <pryzby(at)telsasoft(dot)com> wrote:
>> I think there's an easy answer here that would satisfy everyone; two patches:
>> 0001 to fix the unintentional behavior change;
>> 0002 to reject garbage input: anything with more than 3 dot-separated
>> components, or with 3 components where the first doesn't match
>> current_database.
>>
>> 0001 would be backpatched to v14.
>>
>> If it turns out there's no consensus on 0002, or if it were really hard for
>> some reason, or (more likely) nobody went to the bother to implement it this
>> year, then that's okay.
>
> This might work, but I fear that 0001 would end up being substantially
> more complicated than a combined patch that solves both problems
> together.
Here is a WIP patch that restores the old behavior, just so you can eyeball how large it is. (It passes check-world and I've read it over once, but I'm not ready to stand by this as correct quite yet.) I need to add a regression test to make sure this behavior is not accidentally changed in the future, and will repost after doing so.