[ANSI-Smalltalk] #peekFor:

[Damien Cassou]

I noticed that too when doing Nile. It seems all implementations
decided to implement #peekFor: the same way, ignoring the standard. I
think the standard behavior is useless here because it is equivalent
to "stream peek = anObject" which is clearer that way.

On Wed, Oct 1, 2008 at 3:35 AM, Richard O'Keefe <ok at cs.otago.ac.nz> wrote:
> In the Smalltalk standard, we read
>
> 5.9.2.8 Message: peekFor: anObject
>  Synopsis
>    Peeks at the next object in the stream
>    and returns true if it matches the argument,
>    and false if not.
>  Definition: <gettableStream>
>    Returns the result of sending #= to the first object
>    in the receiver's future sequence values with anObject
>    as the argument.  Returns false if the receiver has
>    no future sequence values.
>  Parameters
>    anObject <Object> uncaptured
>  Return Value
>    <boolean> unspecified
>  Errors
>    none
>
> There is no faintest suggestion here that the state of the
> stream changes in any way.  In particular, there is no
> suggestion that a successful match will result in the stream
> being advanced.
>
> However, the Blue Book says on page 199
>
>        Determine the response to the message [#]peek.
>        If it is the same as the argument, anObject,
>        THEN INCREMENT THE POSITION REFERENCE and
>        answer true.  Otherwise answer false and do
>        not change the position reference.
>
> The comment in VisualWorks says
>
>        Answer false and do not move the position
>        if self next ~= anObject or if the receiver
>        is at the end.  Answer true and
>        INCREMENT POSITION if self next = anObject.
>
> The comment in Squeak says
>
>        Answer false and do not move over the next element
>        if it is not equal to the argument, anObject, or
>        if the receiver is at the end.  Answer true and
>        INCREMENT THE POSITION for accessing elements,
>        if the next element is equal to anObject.
>
> When I wrote my stream classes, I took the standard text literally,
> and my #peekFor: never advanced the stream position.  Whoops!
> I propose the following replacement for the definition:
>
>  Definition: <gettableStream>
>    If the receiver has no future sequence values,
>    returns false.  Otherwise, determines the result of
>    sending #= to the first object in the receiver's
>    future sequence values with anObject as the argument.
>    If that is false, return false with no change to the
>    receiver's state.  If #= answers true, the first
>    object is removed from the receiver's future sequence
>    values and appended to the end of the receiver's past
>    sequence values, then true is returned.
>
>
>
>
>
> _______________________________________________
> ANSI-Smalltalk mailing list
> ANSI-Smalltalk at lists.openskills.org
> http://lists.openskills.org/cgi-bin/mailman/listinfo/ansi-smalltalk
>



-- 
Damien Cassou
Peter von der Ahé: «I'm beginning to see why Gilad wished us good
luck». (http://blogs.sun.com/ahe/entry/override_snafu)