Maciej Sieczka pisze:
Richard Huxton pisze:
1. Add a proxy in front of PostgreSQL and have it intercept the CREATE
TABLE statements and follow them with the correct GRANT statements.
I'd probably modify the code for pgpool/pgbouncer (http://pgfoundry.org/)
That should do it seems! T
Richard Huxton pisze:
Maciej Sieczka wrote:
In PG experts
opinion, could/should a PostGIS editing software lock table for
writing to prevent concurent editing?
The application should either:
1. Take an advisory lock (see the functions/admin functions chapter) so
that it can use another
Richard Huxton pisze:
Maciej Sieczka wrote:
Richard Huxton pisze:
Afraid not. You can't add a trigger to a system table.
Too bad, but thanks much for clarifying this.
The only real solution is to have your own function do both - create
the table and then grant relevant permissions.
Richard Huxton pisze:
Maciej Sieczka wrote:
I need to modify this default PostgreSQL's behaviour, so that the ACL on
a new table in this schema is set to SELECT, INSERT, UPDATE, DELETE,
REFERENCES for "editors", and SELECT for "viewers", without having to
manually GRA
Hello,
I have asked the question on novice ML but it seems there is not much
traffic there. I hope nobody minds I'm re-asking here, on a much more
active list. The problem is as follows:
There are 2 role groups in my cluster: "editors" and "viewers". In each
group there are several users.
In
