From: | Merlin Moncure <mmoncure(at)gmail(dot)com> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | Brendan Jurd <direvus(at)gmail(dot)com>, Andrew Dunstan <andrew(at)dunslane(dot)net>, Florian Pflug <fgp(at)phlo(dot)org>, Robert Haas <robertmhaas(at)gmail(dot)com>, Josh Berkus <josh(at)agliodbs(dot)com>, pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: [PATCH] Exorcise "zero-dimensional" arrays (Was: Re: Should array_length() Return NULL) |
Date: | 2013-04-04 15:10:38 |
Message-ID: | CAHyXU0y6EJ6v1Z0ak=ckpZqHbHOkKSn82E81YOJjxfRTRprp_g@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Wed, Apr 3, 2013 at 11:11 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> In any case, the whole exercise is pointless if we don't change the
> visible behavior of array_dims et al. So I think the idea that this
> would be without visible consequence is silly. What's up for argument
> is just how much incompatibility is acceptable.
The only reasonable answer for this (a provably used, non-security,
non-standards violating, non-gross functionality breakage case) is
*zero*. Our historically cavalier attitude towards compatibility
breakage has been an immense disservice to our users and encourages
very bad upgrade habits and is, IMNSHO, embarrassing.
Changing the way array_dims works for a minor functionality
enhancement is gratuitous and should be done, if at all, via a loudly
advertised deprecation/replacement cycle with a guarding GUC (yes, I
hate them too, but not nearly as much as the expense of qualifying
vast code bases against random compatibility breakages every release).
merlin
From | Date | Subject | |
---|---|---|---|
Next Message | Robert Haas | 2013-04-04 15:56:36 | Re: Multi-pass planner |
Previous Message | Rodrigo Barboza | 2013-04-04 14:59:52 | Why there is a PG_GETARG_UINT32 and PG_RETURN_UINT32? |