Re: Bug in iCalendar export

[Nick Dokos <nicholas.dokos@hp.com>]

Paul R <paul.r.ml@gmail.com> wrote:

> Nick Dokos <nicholas.dokos@hp.com> writes:
> 
> You have some faulty site-lisp code provided by debian emacs packages.
> Because it was the Nth bug I had to track down due to the way debian
> handles site-lisp, I just decided to get rid, for good, of debian
> emacs packages, and do everything by hand. So I'm sorry to tell you
> that I can't tell what line is faulty this time. But I can tell, for
> sure, that it lives in debian site-lisp. Good luck.
> 
> > There seems to be a bug somewhere in org-export-icalendar code where the
> > *ical-tmp* buffer seems to get deleted prematurely, but  I can't figure out
> > where exactly, so I'm hoping that a) it's reproducible and b) somebody will
> > see what's wrong.
> >
> > ...

Paul,

that's an interesting perspective that hadn't crossed my mind at all.
Thanks!

After some testing (it took me a while to get back to it - sorry for the
delay), it certainly looks like the bug is somewhere other than
org-export-icalendar, but I'm not sure that debian site-lisp is to blame
in this case (although I may very well be missing something). I
originally tried this on GNU Emacs 22.1.50.2, which I had gotten from
CVS and built myself. I don't think I use any of Ubuntu's emacs setup at
all. I also tried it on emacs 23.0.60.1 (also obtained from CVS) and
starting it from the build directory with -Q, using whatever version of
org-mode is distributed with that version: I get the error there as
well. So it seems to me that the bug is somewhere fairly deep in emacs
itself (btw, I sent a bug report off to the emacs maintainers - we'll
see what they say). I also tried on an ancient RHEL4.1(?)  system with
the stock emacs 21 version that was packaged with it: I had to do some
work to get org-mode installed there, but I could not then reproduce the
error.

Thanks again for opening my eyes (and thanks to Adam as well),

Nick