For the typical case of ASCII logfile, there is something you should know. This kind of log files often have a 'switch log' mechanism, when the file
becomes to large or after a given timeout the current version is switched and replaced by an empty new one. Generally the old version is archived.
When managing ASCII log, what is interesting is to keep an handle on the last version of the log, not on an archive one, so kazimir has a mechanism
for log switch detection based on the inode number. After every event check pass, it looks for the inode number of the ASCII log. If the result
is different from the one obtain one the previous pass, or if the log could not be opened before (it was not there), then it Troyes to re-open the
log file. So if the log switches, then kazimir will always keeps an handle to the latest version as long as the log name does not change.
