> > however I have an hour or so now so I'll take a quick look at the
> > kill() issue.
>
> OK, quick (untested) hack below. Open to suggestions if
> anyone's got a better way to do it...
>
> Regards, Dave
>
> Index: kill.c
> ===================================================================
> RCS file: /projects/cvsroot/pgsql-server/src/port/kill.c,v
> retrieving revision 1.2
> diff -u -r1.2 kill.c
> --- kill.c 24 Jun 2004 18:53:48 -0000 1.2
> +++ kill.c 24 Aug 2004 15:56:03 -0000
> @@ -25,8 +25,25 @@
> BYTE sigData = sig;
> BYTE sigRet = 0;
> DWORD bytes;
> + DWORD handle;
>
> - if (sig >= PG_SIGNAL_COUNT || sig <= 0)
> + /* Look, but don't touch... */
> + if (sig == 0)
> + {
> + handle = OpenProcess(PROCESS_QUERY_INFORMATION, FALSE,
> pid);
> + if (!handle)
> + {
> + errno = EINVAL;
> + return -1;
> + }
> + else
> + {
> + CloseHandle(handle);
> + return 0;
> + }
> + }
> +
> + /* Regular kill... */
> + if (sig >= PG_SIGNAL_COUNT || sig < 0)
> {
> errno = EINVAL;
> return -1;
>
I think it's better to call uor named pipe. That way we'll find out if
it's actually a postmaster of ours or a different process alltogether
there.
Just make sure the backend can deal with signal number = 0 and discard
that one (haven't got the code around to check right now, but it should
be a simple if-then-don't in the backend code, if it's not already
there. backend/port/win32/signal.c IIRC)
//Magnus
