[Caml-list] Absolute path in the module system

[Caml-list] Absolute path in the module system

[bob zhang]

Dear List,
    Is there a way to express the absolute module path?
    For example,
       Pervasives.(+) --> actually I want it to be absolute Pervasives
module, not shadowed by any 'open' statements?
    Is there a way to do that? if not, is there any deep reason for
not doing this?

Thanks

-- 
Regards
-- Bob

Re: [Caml-list] Absolute path in the module system

[Gabriel Scherer]

As far as I know, the answer is that you can't do that. In fact I
don't even know how you would do that: there is a tension between
powerful naming constructs and stability of references, and I don't
know how you would provide global references in a satisfying way.

This is a classic problem of hygienic macro systems. They are supposed
to protect two different aspects of bound identifiers "hygiene":
- new definition inside macros do not shadow identifiers in scope at
macro invocation/expansion time (it is possible, but not simple, to
emulate this property with careful preservation of scope during
macro-expansion)
- identifiers referenced by macro definitions are not shadowed by
identifier added in scope after that, before macro
invocation/expansion time; this corresponds to your question and I
don't think how you would solve that with a static expansion scheme.

If I remember correctly, the solution of hygienic macro systems is, in
effect, to do a preprocessing step of the pre-expansion source, to
enforce the convention that now name is bound twice in the same scope
by appending unique suffixes to identifiers.

On Tue, Dec 4, 2012 at 9:43 PM, bob zhang <bobzhang1988@gmail.com> wrote:
> Dear List,
>     Is there a way to express the absolute module path?
>     For example,
>        Pervasives.(+) --> actually I want it to be absolute Pervasives
> module, not shadowed by any 'open' statements?
>     Is there a way to do that? if not, is there any deep reason for
> not doing this?
>
> Thanks
>
> --
> Regards
> -- Bob
>
> --
> Caml-list mailing list.  Subscription management and archives:
> https://sympa.inria.fr/sympa/arc/caml-list
> Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
> Bug reports: http://caml.inria.fr/bin/caml-bugs

Re: [Caml-list] Absolute path in the module system

[bob zhang]

Hi,
On Tue, Dec 4, 2012 at 5:59 PM, Gabriel Scherer
<gabriel.scherer@gmail.com> wrote:
> - identifiers referenced by macro definitions are not shadowed by
> identifier added in scope after that, before macro
> invocation/expansion time; this corresponds to your question and I
> don't think how you would solve that with a static expansion scheme.
Yes, that's my concern
> If I remember correctly, the solution of hygienic macro systems is, in
> effect, to do a preprocessing step of the pre-expansion source, to
> enforce the convention that now name is bound twice in the same scope
> by appending unique suffixes to identifiers.
Can you elaborate on that?

I don't see the problems with absolute module path, one simple
solution is to delete the grammar
of 'open statement and local open statements', then for a single
compilation unit, I can figure out the
precise module path here, but this is not  a beautiful solution.

Thanks
>
> On Tue, Dec 4, 2012 at 9:43 PM, bob zhang <bobzhang1988@gmail.com> wrote:
>> Dear List,
>>     Is there a way to express the absolute module path?
>>     For example,
>>        Pervasives.(+) --> actually I want it to be absolute Pervasives
>> module, not shadowed by any 'open' statements?
>>     Is there a way to do that? if not, is there any deep reason for
>> not doing this?
>>
>> Thanks
>>
>> --
>> Regards
>> -- Bob
>>
>> --
>> Caml-list mailing list.  Subscription management and archives:
>> https://sympa.inria.fr/sympa/arc/caml-list
>> Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
>> Bug reports: http://caml.inria.fr/bin/caml-bugs



-- 
Regards
-- Bob