martin, thanks for the details report.  i'm attaching a patch and
ChangeLog - they includes fixes for some other people's changes, as
well.  details below, inline.

On Jan 19, 2008 12:24 PM, martin rudalics <rudalics@gmx.at> wrote:
> To reproduce with Emacs -Q: Open src/keyboard.c, do M-x allout-mode, and
> M-x allout-hide-current-leaves.  Gets me:
>
> Debugger entered--Lisp error: (wrong-type-argument number-or-marker-p nil)
>   -(nil nil 2)
>   (max 1 (- allout-recent-prefix-end allout-recent-prefix-beginning allout-header-subtraction))
>   (setq allout-recent-prefix-end (or (match-end 1) (match-end 2)) allout-recent-prefix-beginning (or (match-beginning 1) (match-beginning 2)) allout-recent-depth (max 1 (- allout-recent-prefix-end allout-recent-prefix-beginning
> allout-header-subtraction)))
>   allout-prefix-data()
> [...]
>   execute-extended-command(nil)
>   call-interactively(execute-extended-command)
>
> Note that `allout-regexp' is
>
> "\\(/\\*_[      ]*[\\.,\\*\\+--=>()\\[{}&!\\?#%\"X@\\$~_\\\\:;\\^/\\*_]\\)\\|\\(/\\*_\\|/\\*_\\)+ ?[^/*_]\\|\f"
>
> hence it cannot possibly match _two_ parenthetical subexpressions.

not quite.  that's more clear if you look at the expression as a
concatenation of a few:

(concat  "\\(/\\*_[
]*[\\.,\\*\\+--=>()\\[{}&!\\?#%\"X@\\$~_\\\\:;\\^/\\*_]\\)"
             "\\|\\(/\\*_\\|/\\*_\\)+ ?[^/*_]"
             "\\|\f")

the actual problem here is that the match on the third alternative,
the ^L formfeed, is not itself a top-level alternative in the
expression.  i've rectified that, and some related things.

along the way i discovered that some error handling code was mangled,
both in a misguided blanket alteration of expressions that look like
error functions (for which `error' condition handlers in
condition-case statements can be mistaken), and in the subsequent
attempt to repair it, which left inert nonsensical code in its wake.


-- 
ken
http://myriadicity.net