Re: [O] Bug: void-function org-list-struct-fix-item-end [7.7 (release_7.7.396.g25c21)]

[Nick Dokos]

Dave Abrahams <address@hidden> wrote:

> > I can't reproduce it.
> 
> Me neither.  But it happened, and I figured I'd better report it.
> 
> >> Package: Org-mode version 7.7 (release_7.7.396.g25c21)
> >
> > This release doesn't exist yet. Latest is about
> > (release_7.7.349.g3d835).
> >
> > Could you make sure your Org is clean first, and then try to reproduce
> > the problem?
> 
> Wouldn't make any difference, since I can't reproduce it at all.
> 
> > Note: the function your Org can't find has been introduced quite
> > recently.
> 
> It went away when I did `M-x load-library org-list' so I'm guessing
> there's some code path that fails to (require org-list) when needed.
> 

More likely it was an artifact of how exactly you upgraded.

A possible scenario: after you pull, you do (load-library "org.el")
which picks up the new version of org.el. Now org.el requires org-list
but that does not mean that the new version of org-list gets loaded: all
that require does is check if the feature is a member of some list
(which it is, since you were using the feature before the pull). The new
org.el uses the org-list-struct-fix-item-end function, but the function
is not defined, since the old org-list.el is still in force.  When you
(load-library "org-list.el") you get the new org-list.el with the
definition of the new function and everything is consistent. Also when
you restart, you get new versions of everything and everything is
consistent.

If you compile org, that creates even more ways that things can get
inconsistent.

Moral of the story: after you pull (and possibly ``make clean; make''),
do M-x org-reload, rather than loading individual files. Even so, if you
see strange things happening, restart your emacs -- or, at the very
least, start a new ``emacs -q -l minimal.emacs'' and test the
strangeness there: chances are that it won't happen, so you'd know you
are in some twisted maze like the one described above.

Nick