| From: | "Andrew Dunstan" <andrew(at)dunslane(dot)net> |
|---|---|
| To: | "PostgreSQL Hackers Mailing List" <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | limiting parallelism in "make check" |
| Date: | 2003-10-28 06:44:08 |
| Message-ID: | 008801c39d1e$e36da7e0$6401a8c0@DUNSLANE |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
I recently posted a patch to pg_regress to limit parallelism for cygwin to a maximum of 10, so that "make check" could succeed. Tom preferred that this should be settable by the user explicitly rather than hard coded (and hidden), and not limited by platform, so that you could say
make MAX_CONNECTIONS=10 check
or
pg_regress --max-connections=10
Adding these switches should be quite straightforward., and I'm prepared to do it. I think it's important that we can run "make check" everywhere.
A slightly simpler alternative would be to provide "make serialcheck" which would run the serial schedule of tests in a temp installation.
Thoughts?
andrew
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Manfred Koizar | 2003-10-28 11:38:53 | Re: Defaults for GUC variables (was Re: pg_ctl reports succes when start fails) |
| Previous Message | Christopher Kings-Lynne | 2003-10-28 06:25:44 | Re: Slightly inconsistent behaviour in regproc? |