[Openmcl-devel] Compiler warnings

[Erik Huelsmann]

On Wed, Oct 21, 2009 at 11:15 AM, Tim Bradshaw <tfb at tfeb.org> wrote:
> On 19 Oct 2009, at 22:52, james anderson wrote:
>
>> is there some use case which argues against such strictness?
>
> Yes.  It's really a pain to use such implementations.  I want a top-
> level which just lets me store things in ad-hoc variables.  I
> *definitely* do not want that to involve declaring a whole load of
> global special variables (usually with names like "X" and "Y" which I
> also use in function arguments and so on, and which thus cause obscure
> and hard-to-reproduce bugs in code).
>
> I use Lisp because I want an environment which helps me, not one which
> gets in my way.  Having to declare everything in a conversatational
> environment is just a huge pain.

Completely agree: you don't want too many side effects to statements
evaluated in the top-level, apart from the immediately intended ones.
Especially if you're working in a REPL of a lisp image which is
running an application which is going to be rolled out using the
save-image method.

Bye,

Erik.