Inspecting functions returning complicated values

Inspecting functions returning complicated values

[Marcin Borkowski]

Hi list,

assume that I want to inspect a function whose return value is a long,
deeply nested list.  I found this way to present it in a readable way:

M-: (setq blah-blah (function-returning-hairy-stuff))
C-h C-v blah-blah

This seems a bit awkward, however.  Is there a better/cleaner way?

TIA,

-- 
Marcin Borkowski
http://octd.wmi.amu.edu.pl/en/Marcin_Borkowski
Adam Mickiewicz University

RE: Inspecting functions returning complicated values

[Drew Adams]

> assume that I want to inspect a function whose return value is a
> long, deeply nested list.  I found this way to present it in a
> readable way:
> 
> M-: (setq blah-blah (function-returning-hairy-stuff))
> C-h C-v blah-blah
> 
> This seems a bit awkward, however.  Is there a better/cleaner way?

M-x pp-eval-expression

Re: Inspecting functions returning complicated values

[Eric Abrahamsen]

Marcin Borkowski <mbork@wmi.amu.edu.pl> writes:

> Hi list,
>
> assume that I want to inspect a function whose return value is a long,
> deeply nested list.  I found this way to present it in a readable way:
>
> M-: (setq blah-blah (function-returning-hairy-stuff))
> C-h C-v blah-blah
>
> This seems a bit awkward, however.  Is there a better/cleaner way?
>
> TIA,

It can also be interesting to start an IELM repl and do things to the
return value there.

Re: Inspecting functions returning complicated values

[Emanuel Berg]

Drew Adams <drew.adams@oracle.com> writes:

> M-x pp-eval-expression

What improvement is that?

     (message "hello")
     (eval-expression (message "hello"))
     (pp-eval-expression (message "hello"))

Here, the first two seem to do the same, namely put it
in the echo area for inspection and then you can use
`C-h e'. pp-eval-expression also puts it in *Messages*
but doesn't echo it what I can see.

But I agree with the other posters, IELM, with

    (ielm)

is the best way to do this.

-- 
underground experts united

RE: Inspecting functions returning complicated values

[Drew Adams]

> > M-x pp-eval-expression
> 
> What improvement is that?
>      (message "hello")
>      (eval-expression (message "hello"))
>      (pp-eval-expression (message "hello"))
> 
> Here, the first two seem to do the same, namely put it
> in the echo area for inspection and then you can use
> `C-h e'. pp-eval-expression also puts it in *Messages*
> but doesn't echo it what I can see.

Maybe it's because you don't know what you're doing?
Maybe you didn't read the doc for `pp-eval-expression'?
Hard to guess.  Just what are you trying to accomplish?

`pp-eval-expression' is a function.  As such, it evaluates its
argument.  What does the doc string tell us it does with its
argument?

  Evaluate EXPRESSION and pretty-print its value.
  Also add the value to the front of the list in the
  variable `values'.

So let's see what you did.  Somehow, you evaluated the sexp
`(pp-eval-expression (message "hello"))'.  Perhaps you used
`C-x C-e', or maybe `M-:', to do that?

However you did it, that evaluates the sexp `(message "hello")'
and passes the resulting value to function `pp-eval-expression'.
That resulting value is the string "hello".

So what does `pp-eval-expression' then do with its argument,
"hello"?

It evaluates it!  And then it pretty-prints the result of that
evaluation.

"hello" evaluates to "hello".  OK, whatever.  Since the
pretty-printed representation of string "hello" is "hello",
`pp-eval-expression' prints "hello" in the echo area.
End of story...almost.

And then, depending on how you evaluated
`(pp-eval-expression (message "hello"))', you might see the
value of _that_ sexp echoed (in the echo area).  What is that
value?  `nil'.

`pp-eval-expression' is used for its side effect, not its value.

Why do `C-x C-e' and `M-:' print the value of the sexp they
evaluate?  Ask them.  But it is generally handy that they do.

So in the echo area, and in `*Messages*', two things are printed:
a pretty-print of the string "hello", followed by `nil'.

In addition, you have access to the result of evaluating the
argument sexp (the string "hello"), which is "hello", as the car
of global variable `values': `C-h v values' or `M-: (car values)'.

Now what was it that you _wanted_ to do?

Typically, one uses `pp-eval-expression' to (1) evaluate an
expression and then (2) pretty-print the result of evaluation.

Because it wants to evaluate an expression argument, you need
to pass it an expression, and `pp-eval-expression' of course
evaluates its argument first, before the function body executes.
(It is not a macro or a special form.)

So a typical use of `pp-eval-expression' would be something
like `(pp-eval-expression 'ctl-x-map)' or
`(pp-eval-expression '(nth 2 (nth 2 font-lock-keywords)))', and
the result of pretty-printing would look something like this:

(keymap
 #^[nil nil keymap
	#^^[3 0 pop-global-mark nil list-buffers save-buffers-kill-terminal list-directory eval-last-sexp find-file nil nil indent-rigidly nil kmacro-keymap downcase-region
	      (keymap
	       (108 . set-language-environment)
	       (99 . universal-coding-system-argument)
              ...

or this:

(2
 (let
     ((type
       (get
	(intern-soft
	 (match-string 1))
	'lisp-define-type)))
   (cond
    ((eq type 'var)
     font-lock-variable-name-face)
    ((eq type 'type)
     font-lock-type-face)
    (t font-lock-function-name-face)))
 nil t)

The first of these is the same kind of representation you see if
you use `C-h v ctl-x-map'.  Why?  Because... `describe-variable'
uses function `pp', just like `pp-eval-expression' does.

But again, depending on how you evaluate the `pp-eval-expression'
sexp, you might be fooled.  If you evaluate it interactively then
you might be fooled by the _command_ that you use to evaluate it.

Using `M-:' or `C-x C-e' (without a prefix arg) prints the value
of the evaluated `pp-eval-expression' sexp, right after
`pp-eval-expression' prints the value of its argument expression.
The hand is quicker than the eye.

But lo and behold, `pp-eval-expression' is a command!  And that
means you can execute it using `M-x' or bind it to a key.  What
happens then?

M-x pp-eval-expression RET
Eval: (nth 2 (nth 2 font-lock-keywords)) RET

And the pretty-printed value is shown in buffer `*Pp Eval Output*'.
And a message tells you "See buffer *Pp Eval Output*."

(That message should not end in a period, but let's not nitpick.)

But wait, how come we didn't need to quote the expression?
Because interactively `pp-eval-expression' uses `read--expression',
and that, well,...reads an expression.  In this case there is no
need to evaluate something to provide `pp-eval-expression' with
its expression argument.

> But I agree with the other posters, IELM, with (ielm)
> is the best way to do this.

Good to know.  To do what?

Re: Inspecting functions returning complicated values

[Emanuel Berg]

Marcin Borkowski <mbork@wmi.amu.edu.pl> writes:

> M-: (setq blah-blah (function-returning-hairy-stuff))
> C-h C-v blah-blah
>
> This seems a bit awkward, however. Is there a
> better/cleaner way?

Ha! Interesting question.

I guess I would simply write the above in an Elisp file
(with an .el extension) and then evaluate it. (Or even
insert it into the buffer.)

Otherwise you can just do it as you do (the
eval-expression or M-: part) and then `C-h e' or
view-echo-area-messages to examine the result in the
*Messages* buffer.

-- 
underground experts united

Re: Inspecting functions returning complicated values

[Pascal J. Bourguignon]

Marcin Borkowski <mbork@wmi.amu.edu.pl> writes:

> Hi list,
>
> assume that I want to inspect a function whose return value is a long,
> deeply nested list.  I found this way to present it in a readable way:
>
> M-: (setq blah-blah (function-returning-hairy-stuff))
> C-h C-v blah-blah
>
> This seems a bit awkward, however.  Is there a better/cleaner way?

In most any (writable) buffer, you can type:

   (prog1 (terpri) (pp (function-returning-hairy-stuff))) C-u C-x C-e

This won't work only in buffer where C-x C-e is not bound to
eval-last-sexp. In those buffers you can type C-u M-x eval-last-sexp RET
instead of C-u C-x C-e

You can also use ielm: 

M-x ielm RET
(pp (function-returning-hairy-stuff)) RET
(pp (other-function-returning-hairy-stuff)) RET
;; etc…


-- 
__Pascal Bourguignon__                 http://www.informatimago.com/
“The factory of the future will have only two employees, a man and a
dog. The man will be there to feed the dog. The dog will be there to
keep the man from touching the equipment.” -- Carl Bass CEO Autodesk