Thread: Re: Prevent COPY FREEZE on Foreign tables
my apologies for the top post in the last reply. I hit send too fast :) I also created a CF entry for this https://commitfest.postgresql.org/52/5544/ I don't think we will need to backpatch, unless someone has a different opinion about this. Regards, Sami
On Wed, Feb 05, 2025 at 01:11:44PM -0600, Sami Imseih wrote: > I don't think we will need to backpatch, unless someone has a different > opinion about this. I think this falls into the category of a bug, so the only reason I can see for not back-patching it is that it could lead to somewhat widespread breakage for existing scripts, etc. which are arguably kind-of working today. That seems unlikely, but it'd be good to hear some other opinions on the matter. -- nathan
> so the only reason I can see > for not back-patching it is that it could lead to somewhat widespread > breakage for existing scripts, etc. which are arguably kind-of working > today. That is my thought for not backpatching. It's not breaking the COPY command, it's just not applying an optimization. Let's see what others think. Sami
On Wed, Feb 05, 2025 at 01:36:57PM -0600, Sami Imseih wrote: >> so the only reason I can see >> for not back-patching it is that it could lead to somewhat widespread >> breakage for existing scripts, etc. which are arguably kind-of working >> today. > > That is my thought for not backpatching. It's not breaking the COPY > command, it's just not applying an optimization. > > Let's see what others think. Given we don't have any other reports or opinions, I'm probably just going to apply the patch to v18. That leaves the door open to back-patch later. The next minor release is next week, anyway, and this doesn't seem urgent. I did find one other thread that mentions this behavior [0]. They floated the idea of blocking COPY FREEZE on foreign tables, but the topic fizzled out in indecision. *shrug* [0] https://postgr.es/m/20181219023814.GA6577%40paquier.xyz -- nathan
> Given we don't have any other reports or opinions, I'm probably just going > to apply the patch to v18. That makes sense. Thanks! Sami
On Thu, Feb 06, 2025 at 10:26:09AM -0600, Sami Imseih wrote: > That makes sense. Thanks! Committed. -- nathan