You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@spamassassin.apache.org by jm...@apache.org on 2004/02/11 21:41:20 UTC

svn commit: rev 6625 - incubator/spamassassin/trunk/spamd

Author: jm
Date: Wed Feb 11 12:41:19 2004
New Revision: 6625

Modified:
   incubator/spamassassin/trunk/spamd/spamd.raw
Log:
added reply to FIXME

Modified: incubator/spamassassin/trunk/spamd/spamd.raw
==============================================================================
--- incubator/spamassassin/trunk/spamd/spamd.raw	(original)
+++ incubator/spamassassin/trunk/spamd/spamd.raw	Wed Feb 11 12:41:19 2004
@@ -293,6 +293,7 @@
     defined(setlogsock($log_socket)) || die $!;
     # The next call is required to actually open the socket.
     # FIXME: mss: wouldn't a openlog('spamd', 'cons,pid,ndelay', $log_facility); be better?
+    # jm: no, it's unportable.  if I recall correctly some perls will die
     syslog('debug', "%s", "spamd starting");
   };
   my $err = $@;
@@ -325,7 +326,7 @@
 # The user wants to log to some file -- open it on STDLOG. Falling back to STDERR
 # if opening the file fails.
 elsif ($log_facility eq 'file') {
-  unless(open(STDLOG, ">> $log_file")) {
+  unless(open(STDLOG, ">>$log_file")) {
     warn "failed to open logfile ${log_file}: $!\n" .
          "reporting logs to stderr\n";
     $log_facility = 'stderr';

to openlog() or not to openlog() (was: Re: svn commit: rev 6625 - incubator/spamassassin/trunk/spamd)

Posted by "Malte S. Stretz" <ms...@gmx.net>.
On Wednesday 11 February 2004 21:41 CET jm@apache.org wrote:
> added reply to FIXME

Replying on-list :)

>      # The next call is required to actually open the socket.
>      # FIXME: mss: wouldn't a openlog('spamd', 'cons,pid,ndelay',
> $log_facility); be better?
> # jm: no, it's unportable.  if I recall  correctly some perls will die
> syslog('debug', "%s", "spamd starting"); 

Are you really sure? The whole syslog() code looks like black magic to me 
and spamd does really weird things with metalog, I guess because of the 
missing $ident.

Currently we open the syslog via a syslog() call but do an openlog() on each 
logmsg() call. That looks more than weird.

And 'perldoc Sys::Syslog' doesn't say anything about known bugs like this, 
it just has a big fat sentence "You should use openlog() before calling 
syslog()."

Cheers,
Malte

-- 
[SGT] Simon G. Tatham: "How to Report Bugs Effectively"
      <http://www.chiark.greenend.org.uk/~sgtatham/bugs.html>
[ESR] Eric S. Raymond: "How To Ask Questions The Smart Way"
      <http://www.catb.org/~esr/faqs/smart-questions.html>