Re: The gnat binder (was: Re: Aerospace Industry says Drop Ada Mandate)

[dennison@romulus23.DAB.GE.COM (Ted Dennison)]

In article <34a2et$9lq@info.epfl.ch>, weber@lglsun.epfl.ch (Mats Weber) writes:
|> In article <347roa$8ob@gnat.cs.nyu.edu>
|> dewar@cs.nyu.edu (Robert Dewar) writes:
|> 
|> > As for Verdix compiling new sources in the right order, I am willing to bet
|> > that it does NOT properly analyze pragma inline cases. In particular, as
|> > I mentioned, there are cases where no order of compilation works in a
|> > traditional system. Consider:
|> 
|> In some situations, it (Verdix) fails when analyzing dependencies on
|> generics (no inlines involved). I often have to tell the compiler to
|> first make the generic, and then its instantiations, otherwise the
|> compiler crashes with an internal error.

I have found this as well. Often it crashes with an internal error only
after filling up the entire disk (which can take quite a while).

T.E.D.