is this a bug?

Francisco Vides Fernandez

unread,

Apr 1, 2003, 1:35:47 PM4/1/03

to

Hello. I have the following:

>
> (define alist '(("a" . 1) ("b" . 2) ("c" . 3)))
> (define anentry (assoc "b" alist))
> anentry
'("b" . 2)
> (set-cdr! y 3)

Error: exception
(set-cdr! '("b" . 2) 3)
1>

But the following works fine

>
> (define alist (list (cons "a" 1) (cons "b" 2) (cons "c" 3)))
> (set! y (assoc "b" alist))
> y
'("b" . 2)
> (set-cdr! y 3)
> y
'("b" . 3)
> alist
'(("a" . 1) ("b" . 3) ("c" . 3))
>

I'm using scsh 0.6.2. I have tested this with other interpreters and
both work fine. Is this a known bug? Am I doing something wrong?

Saludos

Francisco Vides Fernandez

unread,

Apr 1, 2003, 1:42:02 PM4/1/03

to

Sorry, my previous example was wrong. I meant this:

>
> (define alist '(("a" . 1) ("b" . 2) ("c" . 3)))
> (define anentry (assoc "b" alist))
> anentry
'("b" . 2)

> (set-cdr! anentry 3)

Error: exception
(set-cdr! '("b" . 2) 3)
1>

But the following works fine

>
> (define alist (list (cons "a" 1) (cons "b" 2) (cons "c" 3)))
> (set! y (assoc "b" alist))
> anentry
'("b" . 2)

> (set-cdr! anentry 3)
> anentry

RT Happe

unread,

Apr 1, 2003, 1:55:21 PM4/1/03

to

On Tue, 1 Apr 2003, Francisco Vides Fernandez wrote:

> Error: exception
> (set-cdr! '("b" . 2) 3)
>

> But the following works fine

[...]

> I'm using scsh 0.6.2. I have tested this with other interpreters and
> both work fine. Is this a known bug? Am I doing something wrong?

In the first case, you are trying to mutate a constant object. The Scheme
report says <http://www.swiss.ai.mit.edu/~jaffer/r5rs_5.html#SEC22>

In many systems it is desirable for constants (i.e. the values of
literal expressions) to reside in read-only-memory. To express this, it
is convenient to imagine that every object that denotes locations is
associated with a flag telling whether that object is mutable or
immutable. In such systems literal constants and the strings returned by
symbol->string are immutable objects, while all objects created by the
other procedures listed in this report are mutable. It is an error to
attempt to store a new value into a location that is denoted by an
immutable object. (R5RS section 3.4 Storage Model)

See also the section on literal expressions.

rthappe

Francisco Vides Fernandez

unread,

Apr 1, 2003, 2:06:26 PM4/1/03

to

Ok, thanks a lot!

>>>>> "RT" == RT Happe <rth...@web.de> writes:

RT> On Tue, 1 Apr 2003, Francisco Vides Fernandez wrote:
>> Error: exception (set-cdr! '("b" . 2) 3)
>>
>> But the following works fine

RT> [...]

>> I'm using scsh 0.6.2. I have tested this with other interpreters
>> and both work fine. Is this a known bug? Am I doing something
>> wrong?

RT> In the first case, you are trying to mutate a constant object.
RT> The Scheme report says
RT> <http://www.swiss.ai.mit.edu/~jaffer/r5rs_5.html#SEC22>

RT> In many systems it is desirable for constants (i.e. the values
RT> of literal expressions) to reside in read-only-memory. To express
RT> this, it is convenient to imagine that every object that denotes
RT> locations is associated with a flag telling whether that object
RT> is mutable or immutable. In such systems literal constants and
RT> the strings returned by

symbol-> string are immutable objects, while all objects created by

symbol-> the
RT> other procedures listed in this report are mutable. It is an
RT> error to attempt to store a new value into a location that is
RT> denoted by an immutable object. (R5RS section 3.4 Storage Model)

RT> See also the section on literal expressions.

RT> rthappe