gobo-eiffel-develop

[gobo-eiffel-develop] Comment in {DS_LINEAR}.after misleading?

[Stefan S. <sie...@st...>]

Hi,

The description of {DS_LINEAR}.after fooled me this morning ;) Andreas Leitner 
suggested to contact the gobo mailing list, so this is what I'll do now...

Here's {DS_LINEAR}.after:

	after: BOOLEAN is
			-- Is there no valid position to right of internal cursor?
		do
			Result := cursor_after (internal_cursor)
		end

The description "Is there no valid position to right of internal cursor?" made 
me think that 'after' is true whenever the cursor is on the _last_ item of 
the linear (and not beyond). I assumed that "no valid position to the right" 
means that the next position will not point to an item of the container 
anymore.

I don't know if I can propose any better description, as I'm no library 
designer and I believe that you had your reasons to choose this description. 
All i can say is that I expected something more in the style of "Is the 
internal cursor on the position after the last item?". Unfortunately this 
doesn't contain the information that the position right to the cursor is 
invalid, which is quite important, too.

Regards,

Stefan

Re: [gobo-eiffel-develop] Comment in {DS_LINEAR}.after misleading?

[Colin A. <col...@ho...>]

>From: Stefan Sieber <sie...@st...>

>
>Hi,
>
>The description of {DS_LINEAR}.after fooled me this morning ;) Andreas 
>Leitner
>suggested to contact the gobo mailing list, so this is what I'll do now...
>
>Here's {DS_LINEAR}.after:
>
>	after: BOOLEAN is
>			-- Is there no valid position to right of internal cursor?
>		do
>			Result := cursor_after (internal_cursor)
>		end
>
>The description "Is there no valid position to right of internal cursor?" 
>made
>me think that 'after' is true whenever the cursor is on the _last_ item of
>the linear (and not beyond). I assumed that "no valid position to the 
>right"
>means that the next position will not point to an item of the container
>anymore.

That assumption is incorrect. The internal cursor has a valid position after 
the last item - the after-life sentinel (my (joke) terminiology).

The point is that when you are on the last item, it is still valid to call 
`forth'. This will move the cursor forward (hence it's name), and the cursor 
has to go somewhere. It goes to the after-sentinel position. Now there is no 
longer a valid position, so you may no longer call `forth'.

_________________________________________________________________
The next generation of Hotmail is here! http://www.newhotmail.co.uk

Re: [gobo-eiffel-develop] Comment in {DS_LINEAR}.after misleading?

[Stefan S. <sie...@st...>]

> >The description "Is there no valid position to right of internal cursor?"
> >made
> >me think that 'after' is true whenever the cursor is on the _last_ item of
> >the linear (and not beyond). I assumed that "no valid position to the
> >right"
> >means that the next position will not point to an item of the container
> >anymore.
>
> That assumption is incorrect. The internal cursor has a valid position
> after the last item - the after-life sentinel (my (joke) terminiology).
>
> The point is that when you are on the last item, it is still valid to call
> `forth'. This will move the cursor forward (hence it's name), and the
> cursor has to go somewhere. It goes to the after-sentinel position. Now
> there is no longer a valid position, so you may no longer call `forth'.
Sorry for not being precise enough. Now I know that there's a valid position 
after the last item.
But I was not aware of this when I read the Description of 'after'.  To be 
honest, I was only interested in the items (to iterate through them), so I 
guess this is the reason why I didn't think of the sentinels before and after 
the items.

The description is 100% correct, I just interpreted it in a wrong way. Thus I 
assumed that there will be possibly other people who will get it wrong, too. 
Therefore I thought, it could be worth to think about the description.