[m-dev.] for review: introduction of two new directories

[Fergus Henderson]

On 24-Sep-1998, Zoltan Somogyi <zs at cs.mu.OZ.AU> wrote:
> 
> > Why do you need two new directories rather than just one?
> 
> One for C code and one for Mercury code.
> 
> The Mercury code *could* stay in the standard library, but we already
> disable the inclusion of such modules (e.g. debugger_interface) in the
> library manual, so they are logically distinct.

The fact that we don't document some parts isn't good evidence that
those parts should be in different directories.  For example, we might
have decided not to document `parser' and/or `lexer', but even if we had,
they would still belong in the same directory.

Nevertheless, I agree that the modules that you are talking about are
logically distinct: the reason is that they're logically part of the
debugger rather than part of the standard library.

> Separation has two other benefits: the library shared object is smaller,

It's not immediately clear that this is an advantage;
after all, you'll still need to link with the `browser' and `trace'
libraries anyway, in case some part of the program was compiled
with tracing enabled.  So disk usage and link times could
well go up rather than down.  If this is supposed to be part of the
rationale, then I think the log message should go into a bit
more detail.

-- 
Fergus Henderson <fjh at cs.mu.oz.au>  |  "I have always known that the pursuit
WWW: <http://www.cs.mu.oz.au/~fjh>  |  of excellence is a lethal habit"
PGP: finger fjh at 128.250.37.3        |     -- the last words of T. S. Garp.