Hi John, Alex,

Thank you for your explanations, I think I now understand how sort works:
So, for example, in

: (setq A (3 1 2 4 5 6))
-> (3 1 2 4 5 6)
: (setq B (sort A))
-> (1 2 3 4 5 6)
: A
-> (3 4 5 6)
: B
-> (1 2 3 4 5 6)

The symbol A points to the first cell of the list (3 1 2 3 4 5 6), and each
points to the next cell in the list. When sort is applied, the pointing
order of the cells
is changed so that each cell is in the right order in the list. But, A is
still pointing to
the same cell as before, and if that cell has moved, then A ends up
pointing to the
middle of the list.

In the case above, A points to the cell with CAR 3, and when that cell is
moved, A ends
up pointing to the 3rd element of the list. And so, the list that is built
as the value of A
starts at the 3rd element.

My follow up question is, what happens to those first 2 cells, if they're
not assigned to
anything? Do they just stay there? Or are they deleted, perhaps by the
garbage collector?

I also wanted to apologise. When I reread my first mail I realised it was
arrogant of me to
imply that a feature of the language might be a bug.
You have put a lot of effort into this language, and it shows. Several
design decisions that
first surprised me turned out to be awesome, because they let me write code
that was
shorter, easier to read, and prettier. And because they lead to a more
consistent language.

Thank you for all your effort Alex. I like a lot the language you've made.

On Thu, Dec 15, 2016 at 8:30 AM, Alexander Burger <a...@software-lab.de>

> Hi Bruno,
> > But that might be because with 'by, the operation is not destructive.
> This is correct. 'sort' is destructive, and 'by' is not, because it builds
> a
> fresh, private list.
> > Is this normal operation of picolisp? I assumed its not, because I didn't
> > catch any reference to it in the documentation.
> Every destructive function should be marked as such in the reference. Let
> us
> know when you find a case where this is missing.
> - Alex
> --
> UNSUBSCRIBE: mailto:picolisp@software-lab.de?subject=Unsubscribe

Reply via email to