> I honestly expected the VERSION command to query memory instead of > triggering a read of the files on disk. But then again, I can see where > a read of memory could be wrong since (IIRC) a RELOAD doesn't actually > perform the reload until the next new message comes in. Is that still > the case? > - -- > Regards... Todd
I think any tool that uses clamd and which also provides database version info should get that info from clamd and it should be the version that is currently in core. It can also provide the most recently downloaded version, too, but as a minimum, it should always report what is in memory. If the version query happens to result in the daemon reading in any new database before reporting, then clearly there's no need to report both. It can be argued that it should report the version it will use the next time it is called upon to scan a file, and I'd be happy to know that, too. Any tool that does not use clamd but can report the database version should report only the current disk version of the database. Any failure to load a refreshed database should be logged. In fact this is really all I require. It's never happened on my system so I presume I've not ever gotten a corrupt database that has resulted in a failure, or if it has happened, no logging took place. I'm now sufficiently curious to test this. This is such a low priority in my life that I wouldn't care if this is implemented before the sun burns out. It is not something I need to know to understand the health of my clamAV installation, but I'm fussy about data integrity and accurraccy :) dp _______________________________________________ http://lurker.clamav.net/list/clamav-users.html
