Test your ELisp knowledge

Test your ELisp knowledge

[Stefan Monnier]

OK,

Here's a test for you: without trying to run it, guess what error will
be signaled when you do:

    M-: (their-fun (/ my-x 0)) RET


- Stefan "who failed the test"

RE: [External] : Test your ELisp knowledge

[Drew Adams]

> Here's a test for you: without trying to run it, guess what error will
> be signaled when you do:
> 
>     M-: (their-fun (/ my-x 0)) RET
> 
> 
> - Stefan "who failed the test"

Good one - moi aussi.

Re: Test your ELisp knowledge

[Yuri Khan]

On Fri, 9 Feb 2024 at 09:53, Stefan Monnier <monnier@iro.umontreal.ca> wrote:

> Here's a test for you: without trying to run it

Does putting myself in the shoes of a Lisp interpreter count as running?

> guess what error will
> be signaled when you do:
>
>     M-: (their-fun (/ my-x 0)) RET

Something about their-fun not having been defined first. Because if it
is, my next step depends on whether it’s a function or a macro.

If it’s a function then the next is that my-x is not defined as a
variable. If it’s a macro, it might fail on its own, or decide not to
evaluate its argument as a form, but if it does, then same behavior as
with a function.

Otherwise I guess it comes to the division by zero, and that’s the
real question? I guess the result would depend on the type of the
value of my-x, possibly not failing but yielding a positive infinity,
negative infinity or NaN for floats? Otherwise, for integers, I’d
expect a division by zero error; for non-numeric types, a type error.

Re: Test your ELisp knowledge

[Po Lu]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> OK,
>
> Here's a test for you: without trying to run it, guess what error will
> be signaled when you do:
>
>     M-: (their-fun (/ my-x 0)) RET

`their-fun''s function cell is not bound, so eval_sub cannot decide
whether to macroexpand or to funcall it.  :-)

Re: Test your ELisp knowledge

[Manuel Giraud via Emacs development discussions.]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> OK,
>
> Here's a test for you: without trying to run it, guess what error will
> be signaled when you do:
>
>     M-: (their-fun (/ my-x 0)) RET

I'd say: my-x not defined.
-- 
Manuel Giraud

Re: Test your ELisp knowledge

[tomas]

[-- Attachment #1: Type: text/plain, Size: 508 bytes --]

On Fri, Feb 09, 2024 at 11:52:20AM +0100, Manuel Giraud via Emacs development discussions. wrote:
> Stefan Monnier <monnier@iro.umontreal.ca> writes:
> 
> > OK,
> >
> > Here's a test for you: without trying to run it, guess what error will
> > be signaled when you do:
> >
> >     M-: (their-fun (/ my-x 0)) RET
> 
> I'd say: my-x not defined.

I'd side with Yuri Khan and Po Lu: what if `their-fun' was a macro call,
after all?

Quite possibly I'm failing the test now :-)

Cheers
-- 
t

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

Re: Test your ELisp knowledge

[Joost Kremers]

On Fri, Feb 09 2024, tomas@tuxteam.de wrote:
> [[PGP Signed Part:Undecided]]
> On Fri, Feb 09, 2024 at 11:52:20AM +0100, Manuel Giraud via Emacs development
> discussions. wrote:
>> Stefan Monnier <monnier@iro.umontreal.ca> writes:
>> 
>> > OK,
>> >
>> > Here's a test for you: without trying to run it, guess what error will
>> > be signaled when you do:
>> >
>> >     M-: (their-fun (/ my-x 0)) RET
>> 
>> I'd say: my-x not defined.
>
> I'd side with Yuri Khan and Po Lu: what if `their-fun' was a macro call,
> after all?

I initially thought, "well duh, their-fun is not defined", but then I figured,
if Stefan Monnier is asking, it's probably not so obvious, so I went with the
division-by-zero instead...

Which only goes to show that I never thought about the function vs. macro
thingie... :D


-- 
Joost Kremers
Life has its moments

Re: Test your ELisp knowledge

[Manuel Giraud via Emacs development discussions.]

<tomas@tuxteam.de> writes:

> On Fri, Feb 09, 2024 at 11:52:20AM +0100, Manuel Giraud via Emacs development discussions. wrote:
>> Stefan Monnier <monnier@iro.umontreal.ca> writes:
>> 
>> > OK,
>> >
>> > Here's a test for you: without trying to run it, guess what error will
>> > be signaled when you do:
>> >
>> >     M-: (their-fun (/ my-x 0)) RET
>> 
>> I'd say: my-x not defined.
>
> I'd side with Yuri Khan and Po Lu: what if `their-fun' was a macro call,
> after all?
>
> Quite possibly I'm failing the test now :-)

No you won and I failed ;-) (I should have thought about the "am I a
function or macro?" question)
-- 
Manuel Giraud

Re: Test your ELisp knowledge

[Peter Hull]

On Fri, 9 Feb 2024 at 12:17, Joost Kremers <joostkremers@fastmail.fm> wrote:
> I initially thought, "well duh, their-fun is not defined", but then I figured,
> if Stefan Monnier is asking, it's probably not so obvious, so I went with the
> division-by-zero instead...

Me too but is it just because their-fun comes first in text order, e.g.
M-: (+ (/ my-x 0) (their-fun))
gives "(void-variable my-x)"

Re: Test your ELisp knowledge

[tomas]

[-- Attachment #1: Type: text/plain, Size: 235 bytes --]

On Fri, Feb 09, 2024 at 01:31:44PM +0100, Manuel Giraud wrote:

[...]

> No you won and I failed ;-) (I should have thought about the "am I a
> function or macro?" question)

I'm just running on coattails :-)

Cheers
-- 
t

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

RE: [External] : Re: Test your ELisp knowledge

[Drew Adams]

> Me too but is it just because their-fun comes first in text order, e.g.
> M-: (+ (/ my-x 0) (their-fun))
> gives "(void-variable my-x)"

Yes, the args to `+' are evaluated in that
order.  (Similarly for all Elisp functions,
but not necessarily for macros and special
forms.)

But the error in the original case happens
because a sexp such as (their-fun ...)
could be a function call or a macro, with
an unknown/undefined `their-fun'.  Can't
assume that the args need to be evaluated.

Even though Elisp evaluates all args of a
function first-thing (applicative-order
evaluation), which would check `my-x',
it first has to find out if `their-fun'
is a function.

Functions such as `+' and `/' are known
(even built-in).