On Tue, Feb 23, 1999 at 05:08:57PM +0100, Jos Backus wrote: > On Tue, Feb 23, 1999 at 04:16:26PM +0100, Eivind Eklund wrote: > > > Somehow this strikes me as a Bad Thing... > > > > It _is_ a bad thing. I've been pondering what to do with the > > intrusive invariant checks - make them dependent on > > INTRUSIVE_INVARIANTS, perhaps? > > Depends on how dangerous these invariant violations are, I would think. > Iow, do they justify a panic()?
IMO, any invariant violation justifies a panic(). Otherwise, people would not pay heed to them. Invariant violations are pretty serious. However, my opinion is also that invariant checks should be non-intrusive - ie, they should not change the normal code path, only add extra checks. A couple of the invariants we have modify the behaviour to make it possible to check for things, and this should be separate from the ones that doesn't modify the behaviour beyond adding checks. Eivind. To Unsubscribe: send mail to majord...@freebsd.org with "unsubscribe freebsd-current" in the body of the message
