Re: No longer possible to query catalogs for index capabilities?

Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>:

> FWIW, this thread started on 25-July, less than two weeks ago.

Technically speaking, there was a pgsql-jdbc thread started on May 14:
https://www.postgresql.org/message-id/nh72v6%24582%241%40ger.gmane.org

9.6beta1 was released on May 12

The fact that it wasn't raised
> till more than 6 months after we committed the pg_am changes

This means that nobody was testing compatibility of "postgresql's master
branch with existing third-party clients".
Testing against well-known clients makes sense to catch bugs early.

I've added "build postgresql from master branch" test to the pgjdbc's
regression suite a week ago, so I hope it would highlight issues early
(even before the official postgresql beta is released).

However, pgjdbc tests are executed only for pgjdbc commits, so if there's
no pgjdbc changes, then there is no logic to trigger "try newer postgres
with current pgjdbc".

Ideally, postgresql's regression suite should validate well-known clients
as well.
I've no idea how long would it take to add something to postgresql's
buildfarm, so I just went ahead and created Travis test configuration at
https://github.com/vlsi/postgres
Do you think those Travis changes can be merged to the upstream?

I mean the following:
1) Activate TravisCI integration for https://github.com/postgres/postgres
mirror.
2) Add relevant Travis CI file so it checks postgresql's regression suite,
and the other
3) Add build badge to the readme (link to travis ci build) for simplified
navigation to test results.

Here's the commit:
https://github.com/vlsi/postgres/commit/4841f8bc00b7c6717d91f51c98979ce84b4f7df3
Here's how test results look like: https://travis-ci.org/vlsi/postgres

Vladimir