Re: [Chicken-users] Compilation issue with v4.7.0

[Pedro Henrique Antunes de Oliveira]

Actually no, page 16 defines that. Take a look at it. Maybe, there are
some details there that can help you.

On Fri, Aug 12, 2011 at 1:42 PM, Pedro Henrique Antunes de Oliveira
<address@hidden> wrote:
> Maybe, top level execution is undefined.
>
> I read R5RS once, but not paying total attention, and I don't remember
> it defining how top level execution should work.
>
> On Fri, Aug 12, 2011 at 1:40 PM, Paul Colby <address@hidden> wrote:
>>
>> On Aug 12, 2011, at 9:16 AM, Pedro Henrique Antunes de Oliveira wrote:
>>
>>> I am not sure why this happens, but once I've heard that TOP LEVEL
>>> execution is different from normal execution, inside a lambda
>>> expression for example.
>>>
>>> Try compiling this
>>>
>>> (let ((A #t)
>>>      (B 'undefined))
>>>  (if A
>>>      (set! B 'good)
>>>      (set! B 'bad))
>>>  (print B))
>>>
>>> It runs as expected.
>>>
>>> Your problem is probably related to how top level evaluations are executed.
>>>
>>> If I recall correctly, continuations, at top level, behave differently too.
>>>
>>> On Fri, Aug 12, 2011 at 11:56 AM, Paul Colby <address@hidden> wrote:
>>>> Hi,
>>>> I tried the following file as a prelude to doing conditional compilation 
>>>> and ran into the following snag,
>>>> <snip>
>>
>> I already had tried this workaround but was not be able to do top level 
>> defines
>> within the let making conditional compilation less interesting. It might be 
>> worth mentioning that I
>> also tried my little test program with the -O4 switch and it works as 
>> expected yielding 'good' -On
>> where n is 0, 1, 2, or 3 prints 'undefined' per the scoping rules of scheme. 
>> Clearly, top level
>> execution differences shouldn't depend on the compilers optimization level.
>>
>> Thanks for taking a look.
>>
>> Regards
>>
>>
>