| From: | Vladimir Sitnikov <sitnikov(dot)vladimir(at)gmail(dot)com> |
|---|---|
| To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
| Cc: | Craig Ringer <craig(at)2ndquadrant(dot)com>, Yury Zhuravlev <u(dot)zhuravlev(at)postgrespro(dot)ru>, Fabrízio de Royes Mello <fabriziomello(at)gmail(dot)com>, Pgsql Hackers <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: NOT EXIST for PREPARE |
| Date: | 2016-03-24 15:33:02 |
| Message-ID: | CAB=Je-FM4YjUg9sksEZ8htdau=4qu+xMZounAVm4WtbBbquWbQ@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Tom>If you think that's not a protocol change, you are mistaken. It
Tom>changes a behavior that's specified in the protocol documentation.
Even if it requires documentation, this particular change will work seamlessly
across existing implementations of v3 protocol.
For instance, it would not require to update pgbouncer to support that
__ convention.
In other words, __ convention is transparent to pgbouncer.
Consider Prepare2 kind of message is added. Then it would require to update
virtually every software that talks v3 protocol.
That is why I say that "some kind of __ convention" does not require protocol
version bump, while "adding new message" does require the bump.
Just to be clear: I'm not fond of encoding the answer to the universe
into statement name.
However, I find that "name convention" a smart invention.
Vladimir
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Christian Ullrich | 2016-03-24 15:35:23 | Re: BUG #13854: SSPI authentication failure: wrong realm name used |
| Previous Message | Alvaro Herrera | 2016-03-24 15:31:16 | Re: dealing with extension dependencies that aren't quite 'e' |