Re: Proposal: B_ARRIER (addresses wapbl performance?)

[Robert Elz <kre%munnari.OZ.AU@localhost>]

    Date:        Thu, 30 Oct 2008 02:19:09 -0400 (EDT)
    From:        der Mouse <mouse%Rodents-Montreal.ORG@localhost>
    Message-ID:  <200810300628.CAA08626%Sparkle.Rodents-Montreal.ORG@localhost>

  | There are good places for clever and amusing names.

And the "an" (tmac.an) macro package (sim trac.s and tmac.m) - do
those also qualify as "bad examples" ?

I actually see nothing wrong with this - especially when the names
aren't exposed to the general public (and even when they are, when
the cleverness/cuteness is pretty well disguised as it is with the
macro package names).

I always quite liked -DEBUG - though when I used it it tended to be
as
        #ifdef EBUG
        #define DEBUG EBUG
        #endif

and then the rest of the code tests DEBUG - seeing lots of

        #ifdef EBUG
                ...
        #endif

scattered through code always tended to make me look for ABUG as
well...

B_ARRIER I think goes too far, but only because the _ wrecks it.

B_BARRIER is OK (but boring) - I might pick B_TWEEN, and if another
flag was needed for a similar purpose, it would of course be ...

kre