I'm not a C developer but I would've thought that the multigraph
definition is used to prepend or append your defined tests to a specific
set of tests that use/require multiple graphs to display nicely...Am I
mistaken about it's intent? I trust that it's not working for you since
you had to go through the trouble of modifying it.
From what I can tell from your change is you've overwritten the
defaults, so you'd have to explicitly add them back in.
=G=
-----Original Message-----
From: Trent Melcher [mailto:user-c65e78735b17@xymon.invalid]
Sent: Friday, August 17, 2007 1:20 PM
To: user-ae9b8668bcde@xymon.invalid
Subject: [hobbit] Possible Bug in 4.3.0-0.20070722 snapshot
I found a possible bug in the 4.3.0-0.20070722 snapshot.
In the web/hobbitsvc.c (v 1.80 2007/07/18 21:20:15)
This variable is set:
static char *multigraphs = ",disk,inode,qtree,";
With this set the option in the hobbitcgi.cfg for CGI_SVC_OPTS="
--multi-graphs" never gets used. (This is how I found it, trying to
use --multi-graphs=if_load,if_err)
The if statement in the htmllog.c on line 369 is never hit since
multigraph is never NULL.
htmllog.c,v 1.59 2007/07/21 16:13:18
if (multigraphs == NULL) multigraphs = ",disk,inode,qtree,";
I change the hobbitsvc.c to
static char *multigraphs = NULL;
the if statement in htmllog.c now hits because multigraphs is NULL,
however could I have possibly broken other tests with this change? SO
far I havent found anything broken in my beta deployment of the
snapshot, Also has this been fixed in newer snapshots?
Thanks
Trent