bug#31654: emacs-lisp-intro.texi - backtrace

bug#31654: emacs-lisp-intro.texi - backtrace

[Van L]

Hello.

What I’ll do is collect backtrace as documented and shown in GNU Emacs 26.1, here.

I like how the original extended the common prefix ``eval’’ in the naming.

The newer backtrace result looks ``complicated''.

  #+BEGIN_EXAMPLE
    8cda6f8f (Glenn Morris             2007-09-06  1818) ---------- Buffer: *Backtrace* ----------
    8cda6f8f (Glenn Morris             2007-09-06  1819) Debugger entered--Lisp error: (void-function fill-column)
    8cda6f8f (Glenn Morris             2007-09-06  1820)   (fill-column)
    8cda6f8f (Glenn Morris             2007-09-06  1821)   eval((fill-column))
    8cda6f8f (Glenn Morris             2007-09-06  1822)   eval-last-sexp-1(nil)
    8cda6f8f (Glenn Morris             2007-09-06  1823)   eval-last-sexp(nil)
    8cda6f8f (Glenn Morris             2007-09-06  1824)   call-interactively(eval-last-sexp)
    8cda6f8f (Glenn Morris             2007-09-06  1825) ---------- Buffer: *Backtrace* ----------

    Debugger entered--Lisp error: (void-function fill-column)
      (fill-column)
      eval((fill-column) nil)
      elisp--eval-last-sexp(nil)
      eval-last-sexp(nil)
      funcall-interactively(eval-last-sexp nil)
      call-interactively(eval-last-sexp nil nil)
      command-execute(eval-last-sexp)
  #+END_EXAMPLE

bug#31654: Acknowledgement (emacs-lisp-intro.texi - backtrace)

[Van L]

> GNU bug Tracking System writes:
> 
> If you wish to submit further information on this problem, please
> send it to 31654@debbugs.gnu.org.

  #+BEGIN_EXAMPLE
    8cda6f8f (Glenn Morris             2007-09-06  1863) In GNU Emacs 22, you will create a @file{*Backtrace*} buffer that
    8cda6f8f (Glenn Morris             2007-09-06  1864) says:
    8cda6f8f (Glenn Morris             2007-09-06  1865) 
    8cda6f8f (Glenn Morris             2007-09-06  1866) @smallexample
    8cda6f8f (Glenn Morris             2007-09-06  1867) @group
    8cda6f8f (Glenn Morris             2007-09-06  1868) ---------- Buffer: *Backtrace* ----------
    8cda6f8f (Glenn Morris             2007-09-06  1869) Debugger entered--Lisp error: (void-variable +)
    8cda6f8f (Glenn Morris             2007-09-06  1870)   eval(+)
    8cda6f8f (Glenn Morris             2007-09-06  1871)   eval-last-sexp-1(nil)
    8cda6f8f (Glenn Morris             2007-09-06  1872)   eval-last-sexp(nil)
    8cda6f8f (Glenn Morris             2007-09-06  1873)   call-interactively(eval-last-sexp)
    8cda6f8f (Glenn Morris             2007-09-06  1874) ---------- Buffer: *Backtrace* ----------
    8cda6f8f (Glenn Morris             2007-09-06  1875) @end group
    8cda6f8f (Glenn Morris             2007-09-06  1876) @end smallexample

    Debugger entered--Lisp error: (void-variable +)
      eval(+ nil)
      elisp--eval-last-sexp(nil)
      eval-last-sexp(nil)
      funcall-interactively(eval-last-sexp nil)
      call-interactively(eval-last-sexp nil nil)
      command-execute(eval-last-sexp)

  #+END_EXAMPLE

bug#31654: emacs-lisp-intro.texi - backtrace

[Eli Zaretskii]

> From: Van L <van@scratch.space>
> Date: Wed, 30 May 2018 19:23:51 +1000
> 
> I like how the original extended the common prefix ``eval’’ in the naming.
> 
> The newer backtrace result looks ``complicated''.
> 
>   #+BEGIN_EXAMPLE
>     8cda6f8f (Glenn Morris             2007-09-06  1818) ---------- Buffer: *Backtrace* ----------
>     8cda6f8f (Glenn Morris             2007-09-06  1819) Debugger entered--Lisp error: (void-function fill-column)
>     8cda6f8f (Glenn Morris             2007-09-06  1820)   (fill-column)
>     8cda6f8f (Glenn Morris             2007-09-06  1821)   eval((fill-column))
>     8cda6f8f (Glenn Morris             2007-09-06  1822)   eval-last-sexp-1(nil)
>     8cda6f8f (Glenn Morris             2007-09-06  1823)   eval-last-sexp(nil)
>     8cda6f8f (Glenn Morris             2007-09-06  1824)   call-interactively(eval-last-sexp)
>     8cda6f8f (Glenn Morris             2007-09-06  1825) ---------- Buffer: *Backtrace* ----------
> 
>     Debugger entered--Lisp error: (void-function fill-column)
>       (fill-column)
>       eval((fill-column) nil)
>       elisp--eval-last-sexp(nil)
>       eval-last-sexp(nil)
>       funcall-interactively(eval-last-sexp nil)
>       call-interactively(eval-last-sexp nil nil)
>       command-execute(eval-last-sexp)
>   #+END_EXAMPLE

But it's what Emacs shows nowadays.  I don't see any good reasons to
describe in the manual something that is different from the reality, I
think this could only add to confusion.

Backtraces are usually "ugly" in that way, and developers should get
used to that.

If you or someone else have ideas how to make this "nicer" without
lying to the reader, I'm all ears.

Thanks.

bug#31654: emacs-lisp-intro.texi - backtrace

[Van L]

> On 31 May 2018, at 02:54, Eli Zaretskii <eliz@gnu.org> wrote:
> 
> But it's what Emacs shows nowadays.  I don't see any good reasons to
> describe in the manual something that is different from the reality
> 
> If you or someone else have ideas how to make this "nicer" without
> lying to the reader, I'm all ears.

The ``ignore'' passage for pre-version 22 suggests a one line backtrace for the event of function or variable being ``void’’. Neat.

How about something like a universal argument for level of detail of verbosity? for backtrace.

The default is one line where possible for C-x C-e.

The easy to medium setting has lines having a common beautiful prefix where possible.

The extremely difficult setting has all the ugly offal outpourings for the kitchen sink.

bug#31654: emacs-lisp-intro.texi - backtrace

[Eli Zaretskii]

> From: Van L <van@scratch.space>
> Date: Thu, 31 May 2018 13:00:02 +1000
> Cc: 31654@debbugs.gnu.org
> 
> How about something like a universal argument for level of detail of verbosity? for backtrace.

I don't see why it would be useful.  Backtraces must be as accurate as
possible, showing the exact sequence of calls, so that the user who
debugs the problem could step through the relevant code.  Omitting
parts of a backtrace because they might be "ugly" or "confusing" would
be a disservice, I think.  I have never seen any practical use case
where that would be a good idea.  And I don't think I've seen a
debugger that provides such a feature.

Does anyone see a practical use case for such a feature?

bug#31654: Acknowledgement (emacs-lisp-intro.texi - backtrace)

[Van L]

>> GNU bug Tracking System writes:
>> 
>> If you wish to submit further information on this problem, please
>> send it to 31654@debbugs.gnu.org.

#+BEGIN_EXAMPLE
  8cda6f8f (Glenn Morris             2007-09-06  2134) @smallexample
  8cda6f8f (Glenn Morris             2007-09-06  2135) @group
  8cda6f8f (Glenn Morris             2007-09-06  2136) ---------- Buffer: *Backtrace* ----------
  8cda6f8f (Glenn Morris             2007-09-06  2137) Debugger entered--Lisp error:
  8cda6f8f (Glenn Morris             2007-09-06  2138)          (wrong-type-argument number-or-marker-p hello)
  8cda6f8f (Glenn Morris             2007-09-06  2139)   +(2 hello)
  8cda6f8f (Glenn Morris             2007-09-06  2140)   eval((+ 2 (quote hello)))
  8cda6f8f (Glenn Morris             2007-09-06  2141)   eval-last-sexp-1(nil)
  8cda6f8f (Glenn Morris             2007-09-06  2142)   eval-last-sexp(nil)
  8cda6f8f (Glenn Morris             2007-09-06  2143)   call-interactively(eval-last-sexp)
  8cda6f8f (Glenn Morris             2007-09-06  2144) ---------- Buffer: *Backtrace* ----------
  8cda6f8f (Glenn Morris             2007-09-06  2145) @end group
  8cda6f8f (Glenn Morris             2007-09-06  2146) @end smallexample

  Debugger entered--Lisp error: (wrong-type-argument number-or-marker-p hello)
    +(2 hello)
    eval((+ 2 'hello) nil)
    elisp--eval-last-sexp(nil)
    eval-last-sexp(nil)
    funcall-interactively(eval-last-sexp nil)
    call-interactively(eval-last-sexp nil nil)
    command-execute(eval-last-sexp)

#+END_EXAMPLE

bug#31654: Acknowledgement (emacs-lisp-intro.texi - backtrace)

[Eli Zaretskii]

> From: Van L <van@scratch.space>
> Date: Fri, 1 Jun 2018 23:42:09 +1000
> 
>   Debugger entered--Lisp error: (wrong-type-argument number-or-marker-p hello)
>     +(2 hello)
>     eval((+ 2 'hello) nil)
>     elisp--eval-last-sexp(nil)
>     eval-last-sexp(nil)
>     funcall-interactively(eval-last-sexp nil)
>     call-interactively(eval-last-sexp nil nil)
>     command-execute(eval-last-sexp)

Thanks, fixed.