Re: Object naming conventions (was: Do we need "Mission-Critical" software?)

[Robert Dewar <robert_dewar@my-deja.com>]

In article <9356q9$ig9$1@trog.dera.gov.uk>,
  "Kevin Rigotti" <rigotti@atc.dera.gov.uk> wrote:

> GNAT is well behaved in its choice of names, for example a
> procedure Reset in package Cactus.Random becomes
> "cactus__random__reset". Nothing surprising
> there.

Unfortunately it is not quite that simple. First, overloading
obviously introduces additional problems.

Second, and more subtly, in Ada 95, it is possible to have
two different non-overloadable entities, both of which have
the name a.b.c, for example, you can have two variables in
a program which have this same fully qualified name. It is
an excercise for the reader to figure out how (for the answer
see exp_dbug.ads in the GNAT sources).

So naming conventions are a little tricky. GNAT goes to great
efforts to make the names transparent so that debugging with
non-Ada aware debuggers (such as are likely to be available
in hardware emulators) is reasonably practical (the other
naming scheme would indeed be a bit hostile in this context!).

For full details on the GNAT naming scheme, which also includes
details of the encodings GNAT uses to present full Ada
information using system standard debugging formats, see
the source of Exp_Dbug in file exp_dbug.ads.


Sent via Deja.com
http://www.deja.com/