Re: Using eval with environments and R7RS modules

[Damien MATTEI <damien.mattei@oca.eu>]

yes Per i almost never use EVAL, i finally find a better solution than 
with EVAL in the macro i was writting.

I needed to EVAL a quoted argument of a macro but ,it was finally useless.


Le 30/11/2020 à 13:21, Per Bothner a écrit :
> On 11/30/20 2:13 AM, Damien MATTEI wrote:
>> (define (foo L)
>>
>>     (eval L (interaction-environment))
>>
>> and i got an error like UNBOUND VARIABLE L
>>
>> but this not only in Kawa,it is in Scheme, i was in a R5RS ,
>
> The eval functon is really not compatible with lexical scoping, at least
> not unless you use a different evaluation strategy which make Scheme
> run much slower and use more memory. And that would be all the time,
> not just when eval is used (possibly unless you made eval a special
> syntax the compiler could recognize, rather than a procedure).
>
> Scheme from the very beginning has focused on lexical scoping and
> being able to compile it to efficient code:
>
> https://en.wikisource.org/wiki/Lambda_Papers
>
>> if there is a way to know the current bindings from EVAL i will be happy
>
> Dynamic bindings, yes, but lexical bindings, no.  And that won't change.
>
>> to know it, because without this , i do not see any use for EVAL ,
>
> I've many times said there is very little use for eval, and that too many
> people overuse it.

strange that the SICP cover book put the couple EVAL / APPLY as 
important notion, APPLY seems more usefull,

i will try to find examples in the french edition of this book i have

>
> If you think you have a good use for eval, you're probably wrong.
for now no, but i just discover i have to rewrite code that only works 
in toplevel (Gasp...)
>
>> having only the toplevel bindings is not enought for development.
>
> If you mean that eval is not very useful for debugging, that is true.


Damien