pg_ctl.c

Bruce Momjian wrote:

I am almost finished Andrew's version of pg_ctl.c. Here is the current
version I am using.

Sorry, I should have sent this to patches instead.

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073

Bruce Momjian wrote:

I am almost finished Andrew's version of pg_ctl.c. Here is the current
version I am using.

Bruce,

Thanks for advancing this.

I think this is probably fairly close for Unix - the main thing missing
is a more determined effort to find out what port is actually being used
(i.e we need to check the options and the config file).

For Windows we will need different command lines in start_postgres() (or
if that won't work direct calls to CreateProcess() ), and a way to make
sure we don't collide with management by SCM - probably the easiest way
once we have a way of detecting that this is so is to issue "net
{stop,start} servicename" commands instead of sending
SIGINT/SIGQUIT/SIGTERM followed by a call to run postmaster.

If you can get it in a shape that we can get it in before feature freeze
I will try to do any polishing necessary afterwards - unfortunately
pressure of work is restricting my time right now.

cheers

andrew

Andrew Dunstan wrote:

Bruce Momjian wrote:

I am almost finished Andrew's version of pg_ctl.c. Here is the current
version I am using.

Bruce,

Thanks for advancing this.

I think this is probably fairly close for Unix - the main thing missing
is a more determined effort to find out what port is actually being used
(i.e we need to check the options and the config file).

Oh, I see that now. Let me work on that.

For Windows we will need different command lines in start_postgres() (or
if that won't work direct calls to CreateProcess() ), and a way to make
sure we don't collide with management by SCM - probably the easiest way
once we have a way of detecting that this is so is to issue "net
{stop,start} servicename" commands instead of sending
SIGINT/SIGQUIT/SIGTERM followed by a call to run postmaster.

If you can get it in a shape that we can get it in before feature freeze
I will try to do any polishing necessary afterwards - unfortunately
pressure of work is restricting my time right now.

Agreed. Once it is in CVS we can adjust it.

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073