Nagios folks:
I've been looking at Nagios again recently and you all deserve a lot of
credit for it. It's really come a long way since its NetSaint days!
In the process of testing out Nagios, I noticed that when I would enter
downtimes via the CGI interface that the times would not be what I
entered. It would always be some seemingly random time after what I
entered. (I.e. 17:00 would become 17:36 or some such.)
I quickly discovered that the problem was me-- I neglected to enter 17:00
as "17:00:00" with the seconds included and the CGI was tacking on some
random number of seconds since the value was uninitialized. The best fix
for this would be to have the CGI validate all its inputs, but the next
best fix is to simply initialize all the time values appropriately.
Below is a patch to do just that, should you want to. I regret that since
the SourceForge CVS servers are unavailable that I can't check CVS to see
if this was already fixed (or if not, at least provide a patch versus CVS
rather than the 1.0 release code.)
Since the CVS servers aren't working, here's the patch versus 1.0:
-----
diff -u cmd.c.orig cmd.c
--- cmd.c.orig Sun Nov 10 17:13:09 2002
+++ cmd.c Wed Jan 15 13:40:29 2003
@@ -2251,6 +2251,20 @@
int string_to_time(char *buffer, time_t *t){
struct tm lt;
+ /* Initialize some variables just in case they don't get parsed
+ by the sscanf() call. A better solution is to also check the
+ CGI input for validity, but this should suffice to prevent
+ strange problems if the input is not valid.
+ Jan 15 2003 Steve Bonds */
+ lt.tm_mon = 0;
+ lt.tm_mday = 1;
+ lt.tm_year = 1900;
+ lt.tm_hour = 0;
+ lt.tm_min = 0;
+ lt.tm_sec = 0;
+ lt.tm_wday = 0;
+ lt.tm_yday = 0;
+
sscanf(buffer,"%02d/%02d/%04d
%02d:%02d:%02d",<.tm_mon,<.tm_mday,<
..tm_year,<.tm_hour,<.tm_min,<.tm_sec);
lt.tm_mon--;
-----
-- Steve Bonds
This post was automatically imported from historical nagios-devel mailing list archives
Original poster: r1p6os402@sneakemail.com