Re: EMERGENCY: new remote root exploit in UW imapd

[alec () dakotacom net (Alec Kosky)]

On 17-Jul-98 Craig Spannring wrote:
> C should not be used for trusted programs.  The lack of true arrays
> with array bounds checking alone makes it too hazardous.  How many
> buffer overflow attacks would we hear about if the trusted server
> programs were written using a language with bounds checking like
> Modula-2 or Ada?  Zero.

   I like Ada's super-tight type, although at times it's trying, to say the
least. The only major complaint I have against it is the lack of widespread
support for it. I have only found one *nix-based compiler (GNAT), and I was not
too impressed with it. I haven't used it extensively, so I can't comment on too
much, but from what I remember it didn't have a large set of libraries. Perhaps
things have changed in the past year... On the DOS/Windows based side of
things, the situation is only slightly better (last I knew). The only two
decent (but commercial) compilers that I knew of were the Meridian Ada compiler
and the Janus Ada compiler, and the Meridian was by far the superior. This
brings me to the point: Yes, choosing a language like Ada for secure trusted
programs is to be preferred (although nothing can compensate for poor coding
technique), there is a definite need for more support. What is the current
state of Ada compiler technology looking like? Have things changed much?

--Alec--