logger subprocess

Bruce Momjian wrote:

Your patch has been added to the PostgreSQL unapplied patches list at:

http://momjian.postgresql.org/cgi-bin/pgpatches

It will be applied as soon as one of the PostgreSQL committers reviews
and approves it.

Do not apply.
I'm investigating issues under win32.
Main issue:
pgpipe doesn't create a pipe, but sockets. win32 doesn't like to
redirect stderr into sockets...

There's also an issue about file handles not being inherited. The file
handles are there with _spawnl, but not CreateProcess (despite
bInheritHandles=true). Still hunting.

Regards,
Andreas

Retracted by author.

---------------------------------------------------------------------------

Andreas Pflug wrote:

This is the known patch, with following changes:

- realStdErr handed over for EXEC_BACKEND, but still not tested
- Sometimes EMFILE is received in the logger's process queue, when a
backend ended after a SSL connection was interrupted. This is ignored
now (previously it forced an exit(1) and restart of the subprocess)
- log_destination needs to be PGC_POSTMASTER, because the logger process
creation depends on that.
- no functions included

Regards,
Andreas

---------------------------(end of broadcast)---------------------------
TIP 4: Don't 'kill -9' the postmaster

-- 
  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

This patch is updated to compile under WIN32.

It does *not* implement the syslogger subprocess facility. This will
have to be implemented in completely win32 specific way; see discussion
on win32-hackers.

Regards,
Andreas