21 Sep
2005
21 Sep
'05
11:47 a.m.
On Wed, 2005-09-21 at 01:40 +0200, Roberto Nibali wrote:
This indeed is a bit strange as -HUP should not kill the process per se with regard to syslog-ng. I've just tested this and and if I do it by hand the syslog-ng process is indeed killed by this, although not all the time. Some short time digging reveiled following:
It's actually 100% reproducable, so I spent 1 hour looking at this. I've suspected an internal glib error first but this is not the case. I am now using the newest glib-2.8.1 (earlier using glib-2.2.3), plus I have added some debugging code and some implicit g_string_sized_new() calls:
Thanks for the info. I'm trying to look into this as soon as I get some time. -- Bazsi