bug#16294: 24.3.50; debugger error: "button-at: Args out of range: 0"

bug#16294: 24.3.50; debugger error: "button-at: Args out of range: 0"

[Drew Adams]

In the debugger (`debug', not `edebug'), I get to this:

(cond ((string-match-p "\\`[0-9a-fA-F]+\\'" input)
       (string-to-number input 16))
      ((string-match-p "^#" input)
       (read input))
      (t
       (cddr (assoc-string input mctized-cands t))))

Then, when it tried to evaluate this sexp:
(string-match-p "\\`[0-9a-fA-F]+\\'" input), which I can evaluate with
no problem using `e', the debugger is exited in error, with the
message "button-at: Args out of range: 0".

FWIW, the value of INPUT was this (a propertized string):
#("DEVICE CONTROL FOUR	^T" 0 1 (face nil) 1 21 (face nil))

(The "^T" is actually the single control character Control-T.  Changed
to the string "^T" for the bug report.)

Evaluating it using `e' returns nil, as expected.  Similarly, using `e'
I can evaluate (string-match-p "^#" input), to nil.

If I turn off debugging, the code works fine.  The debugger itself seems
to have a problem with it.

(Var `lexical-binding' is nil, here.  The code uses dynamic scoping.)


In GNU Emacs 24.3.50.1 (i686-pc-mingw32)
 of 2013-12-27 on ODIEONE
Bzr revision: 115778 rgm@gnu.org-20131228000456-1797o8z6veuyozs0
Windowing system distributor `Microsoft Corp.', version 6.1.7601
Configured using:
 `configure --prefix=/c/Devel/emacs/binary --enable-checking=yes,glyphs
 'CFLAGS=-O0 -g3' LDFLAGS=-Lc:/Devel/emacs/lib
 CPPFLAGS=-Ic:/Devel/emacs/include'

bug#16294: 24.3.50; debugger error: "button-at: Args out of range: 0"

[Lars Ingebrigtsen]

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

> In the debugger (`debug', not `edebug'), I get to this:
>
> (cond ((string-match-p "\\`[0-9a-fA-F]+\\'" input)
>        (string-to-number input 16))
>       ((string-match-p "^#" input)
>        (read input))
>       (t
>        (cddr (assoc-string input mctized-cands t))))
>
> Then, when it tried to evaluate this sexp:
> (string-match-p "\\`[0-9a-fA-F]+\\'" input), which I can evaluate with
> no problem using `e', the debugger is exited in error, with the
> message "button-at: Args out of range: 0".

Button-at?  That's weird.

Are you still seeing this problem?

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#16294: 24.3.50; debugger error: "button-at: Args out of range: 0"

[Drew Adams]

> Button-at?  That's weird.

Yes, very.

> Are you still seeing this problem?

I have seen it recently, IIRC, but I have not seen it often.
(The context is the debugger, which I don't use all day long.)

bug#16294: 24.3.50; debugger error: "button-at: Args out of range: 0"

[npostavs]

merge 16294 23949
quit

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

> In the debugger (`debug', not `edebug'), I get to this:
>
> (cond ((string-match-p "\\`[0-9a-fA-F]+\\'" input)
>        (string-to-number input 16))
>       ((string-match-p "^#" input)
>        (read input))
>       (t
>        (cddr (assoc-string input mctized-cands t))))
>
> Then, when it tried to evaluate this sexp:
> (string-match-p "\\`[0-9a-fA-F]+\\'" input), which I can evaluate with
> no problem using `e', the debugger is exited in error, with the
> message "button-at: Args out of range: 0".

This is actually the same bug as 23949, the problem is that
string-match-p let-binds inhibit-changing-match-data and causes problems
for the debugger code.
(http://debbugs.gnu.org/cgi/bugreport.cgi?bug=23949#56)

In version 24.5, emacs -Q then evaluate (string-match-p "." nil) gives
"button-at: Args out of range: 0".  In version 25.0.95 it gives
"find-auto-coding: Wrong type argument: stringp, nil"; presumably the
specifics of the code the debugger runs changed somewhat.