wmperry(a)aventail.com (William M. Perry) writes:
Shenghuo ZHU <zsh(a)cs.rochester.edu> writes:
> When I evaluate
> (url-insert-file-contents "http://www.mail-archive.com/lists.html")
>
> #4 0x4040d868 in abort () at ../sysdeps/generic/abort.c:88
> #5 0x808e7ca in assert_failed () at emacs.c:2825
> #6 0x81331a7 in bufpos_to_bytind_func (buf=0x8852670, x=1025) at insdel.c:581
> #7 0x8133b8a in bufpos_to_bytind (buf=0x8852670, x=1025) at buffer.h:985
> #8 0x80e4cc6 in map_extents (from=1, to=1025,
> fn=0x80ea2f0 <report_extent_modification_mapper>, arg=0xbfffdf78,
> obj=142943856, after=0x0, flags=15360) at extents.c:2272
> #9 0x80f023d in report_extent_modification (buffer=142943856, start=1,
> end=1025, inside=0x8269b34, afterp=1) at extents.c:4697
> #10 0x8135b60 in signal_after_change (buf=0x8852670, start=1, orig_end=1,
> new_end=1025) at insdel.c:2298
Eek, looks like a bug in the after-change-hooks processing...
indeed. We have seen something like this (at least) once before [1].
Have a look at David Byers' (excellent) report at:
http://www.xemacs.org/list-archives/xemacs-beta/9905/msg00350.html
the short recipe is:
(switch-to-buffer "FOO")
(insert "123456")
(defun acf (start end len)
(widen)
(goto-char 4)
(delete-char 1))
(let ((after-change-functions '(acf)))
(save-restriction
(narrow-to-region 4 4)
(transpose-lines 1)))
this instantly triggers the assertion.
I think I looked at this some time ago, but can't remember the
details. I'll try to investigate some more in the next couple of days,
but any help is of course greatly appreciated.
Gunnar
Footnotes:
[1] Actually I immediatedly remembered it, but to be honest it took
me a couple of hours of grepping through my harddisk and
searching the BTS until I finally tried searching in the
xemacs-beta archive and found it :-)