On Tue, Jul 15, 2008 at 9:33 AM, Gabriel Michael Black <[EMAIL PROTECTED]> wrote: > On Mon, 14 Jul 2008, nathan binkert wrote: >> >> I'd argue that I'd prefer to get rid of cassert and create our own >> assert for everything if we're going to start messing with this. This >> also argues that the name assert should change so it is not confused >> with the cassert version. >> >> Am I the only one that thinks this is an issue? > > I think it's a valid concern. Parallel systems that do the similar things > tend to cause problems.
It's worth thinking about, but not over-engineering... we already have two paths that m5 can take on unforeseen errors with assert() and panic(). If our enhanced flavors of assertion all call panic() (I think having them call assert() is too restrictive), then things really aren't that different as far as I can see. I don't see why providing enhanced forms of assert (under different names) should necessarily cause us to get rid of the classic version; I was viewing them more as complements than as replacements. I definitely agree that we should not attempt to replace the classic assert with our own version that has the same name... that way lies madness. If there's a desire to deprecate classic assert once we have our own better alternative, I won't fight it, but I personally don't see the need. I think that decision can be decoupled from whether or not we actually provide those better alternatives. Steve _______________________________________________ m5-dev mailing list m5-dev@m5sim.org http://m5sim.org/mailman/listinfo/m5-dev