> Part of maintaining a virus scanner may well include backporting > things, and that may be a fair bit of work to do. There are
> But that doesn't mean that stability has become less important, it At one point do we decide that the backport may have had a significant impact on stability? Say someone implements a significant new feature in spamassassin to handle a particularly virulent new kind of spam, and this new feature doubles the code size of the source. It's probably possible to rip that new code out of the upstream source and "backport" it into the stable version -- but I'd argue that the resulting binary is not much like the original "stable" one. -- thanks, Will