Re: [PATCH] server_version_num should be GUC_REPORT
| От | Tom Lane |
|---|---|
| Тема | Re: [PATCH] server_version_num should be GUC_REPORT |
| Дата | |
| Msg-id | 5204.1420815207@sss.pgh.pa.us обсуждение |
| Ответ на | [PATCH] server_version_num should be GUC_REPORT (Craig Ringer <craig@2ndquadrant.com>) |
| Ответы |
Re: [PATCH] server_version_num should be GUC_REPORT
|
| Список | pgsql-hackers |
Craig Ringer <craig@2ndquadrant.com> writes:
> While looking into client code that relies on parsing server_version
> instead of checking server_version_num, I was surprised to discover that
> server_version_num isn't sent to the client by the server as part of the
> standard set of parameters reported post-auth.
Why should it be? server_version is what's documented to be sent.
> The attached patch marks server_version_num GUC_REPORT and documents
> that it's reported to the client automatically.
I think this is just a waste of network bandwidth. No client-side code
could safely depend on its being available for many years yet, therefore
they're going to keep using server_version.
regards, tom lane
В списке pgsql-hackers по дате отправления: