Hello,
this review is related to Peter' proposal
http://www.postgresql.org/message-id/50F55F5A.3050207@gmx.net
This patch is relatively simple - just introduce necessary interface
to existent functionality.
1) a proposal was accepted with significant agreement - there was only
one disagreement
"> Thoughts?
In this approach, we cannot change the settings in the standby?
If yes, I don't like this approach.
Regards,
--
Fujii Masao"
This functionality is not new, and we have not to solve hot-standby
necessary - or we can solve possible issues in separate task.
2) This patch was applied cleanly and compilation was without errors
and warnings
3) regress tests
=======================All 135 tests passed.
=======================
4) I checked functionality without unexpected issues
5) Open question
* I think so doc is not fully correct
http://www.postgresql.org/message-id/CAFj8pRBf6suKewDCiXiGy=NeYY_Ns9CAZemomvRYsAQ=UpLzNA@mail.gmail.com
* syntax
when I try some variants I got not user friendly error message
postgres=# alter role set work_mem = '1MB';
ERROR: unrecognized role option "work_mem"
LINE 1: alter role set work_mem = '1MB'; ^
Regards
Pavel