Re: Possible security hole? nnfolder evaluates local variables..

Re: Possible security hole? nnfolder evaluates local variables..

[vroonhof]

> > The problem:
> >   XEmacs evals local variables for nnfolder files
> > 
> > The symptom:
> >   This appeared in the minibuffer
> > 
> > Fetching headers for nnfolder:xemacs-beta...done
> > File local-variables error: (void-function winmgr-mode)
>

> [refering to definition of nnheader-find-file-noselect]
> The `enable-local-variables' (under Emacs, at least) should ensure
> that no variables are evaled.  Is there some other variable under
> XEmacs that controls this?

XEmacs uses the same variable..

> nnfolder uses this to load the folder:> 
> (defun nnheader-find-file-noselect (&rest args)

Ahhh. That is the problem...

Under XEmacs it nnheader-find-file-noselect is shadowed by
nnheader-xemacs-find-file-noselect (from nnheaderxm.el). A large ugly
function that looks like a verbatim copy of find-file-noselect from
the XEmacs files.el. It proably does the enable-local-variables thing
wrong. I am not sure why it calls after-find-file functions at all.

The last reference I can find to it is in the Changelog of Gnus 5.2.39.
where there is a "simplify" entry. I cannot see why it was introduced.

Does anybody know?

-- 
Jan Vroonhof                    http://www.math.ethz.ch/~vroonhof/
Mathematik,                                  vroonhof@math.ethz.ch
HG E16, ETH-Zentrum,                      Tel: +41-1-6325456/25154
Raemistrasse 101, CH-8092 Zuerich.              Fax: +41-1-6321085

Re: Possible security hole? nnfolder evaluates local variables..

[Steven L Baur]

This is one tangled mess.  Are either of you using tm?  It also
contains redefinitions of `nnheader-find-file-noselect' and
`nnheader-insert-file-contents', except that they're correct (sigh,
the correct code only gets applied to XEmacs/Mule).

vroonhof  <vroonhof@math.ethz.ch> writes:

> Under XEmacs it nnheader-find-file-noselect is shadowed by
> nnheader-xemacs-find-file-noselect (from nnheaderxm.el). A large ugly
> function that looks like a verbatim copy of find-file-noselect from
> the XEmacs files.el.

Nope.  It's a copy of an *old* version of find-file-noselect.

> It proably does the enable-local-variables thing wrong.

It looks like it, but I'm not sure what it is trying to accomplish.
-- 
steve@calag.com baur
Unsolicited commercial e-mail will be billed at $250/message.

Re: Possible security hole? nnfolder evaluates local variables..

[Jan Vroonhof]

Steven L Baur <steve@calag.com> writes:

> This is one tangled mess.  Are either of you using tm?  It also
> contains redefinitions of `nnheader-find-file-noselect' and
> `nnheader-insert-file-contents', except that they're correct (sigh,
> the correct code only gets applied to XEmacs/Mule).

I am using tm, but not XEmacs/Mule (19.15 in fact).

> 
> It looks like it, but I'm not sure what it is trying to accomplish.

I hope that larsi can find the old Changelog entry that introduced it.
Going back to rgnus learns that it was introduced BEFORE the
file-local variables thing was introduced in the ordinary version.

Jan

Re: Possible security hole? nnfolder evaluates local variables..

[Lars Magne Ingebrigtsen]

vroonhof@math.ethz.ch writes:

> The last reference I can find to it is in the Changelog of Gnus 5.2.39.
> where there is a "simplify" entry. I cannot see why it was introduced.

5.2.17, you mean.

For the first time since I established the Total Gnus CVS Archive,
I've actually used it for something.  Whee!

The "simplify" was exactly that -- a `boundp' thing was removed, but
there was no change otherwise.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: Possible security hole? nnfolder evaluates local variables..

[Lars Magne Ingebrigtsen]

Jan Vroonhof <vroonhof@math.ethz.ch> writes:

> I hope that larsi can find the old Changelog entry that introduced it.
> Going back to rgnus learns that it was introduced BEFORE the
> file-local variables thing was introduced in the ordinary version.

I have no idea why nnheaderxm.el includes the copies of these
functions at all.  Steven, you're the one who put them back in after I
removed them -- was it an XEmacs 19.13 thing or something?

In comparison, the function used under Emacs is rather small and
straightforward: 

(defun nnheader-find-file-noselect (&rest args)
  (let ((format-alist nil)
	(auto-mode-alist (nnheader-auto-mode-alist))
	(default-major-mode 'fundamental-mode)
	(enable-local-variables nil)
        (after-insert-file-functions nil))
    (apply 'find-file-noselect args)))

Shouldn't this work under XEmacs as well?

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: Possible security hole? nnfolder evaluates local variables..

[Steven L Baur]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> I have no idea why nnheaderxm.el includes the copies of these
> functions at all.  Steven, you're the one who put them back in after I
> removed them -- was it an XEmacs 19.13 thing or something?

I recall that (digging ...) September Gnus v0.27 removed by you,
September Gnus v0.28 restored by me.

Tue Jan 16 17:25:28 1996  Steven L. Baur  <steve@miranova.com>
 ...
	* gnus-xmas.el (insert-file-contents-literally): Restored from
	v0.26 nnheader.el since XEmacs 19.13 doesn't have this function.

This was the period I was working around an extremely ugly XEmacs
internals bug that interacted very badly with nnmh.

If this is the reason those functions are still there, I think it's
safe to kill them now. :-)

> In comparison, the function used under Emacs is rather small and
> straightforward: 

> (defun nnheader-find-file-noselect (&rest args)
>   (let ((format-alist nil)
> 	(auto-mode-alist (nnheader-auto-mode-alist))
> 	(default-major-mode 'fundamental-mode)
> 	(enable-local-variables nil)
>         (after-insert-file-functions nil))
>     (apply 'find-file-noselect args)))

> Shouldn't this work under XEmacs as well?

It looks O.K. to me.
-- 
steve@calag.com baur
Unsolicited commercial e-mail will be billed at $250/message.

Re: Possible security hole? nnfolder evaluates local variables..

[Hrvoje Niksic]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> (defun nnheader-find-file-noselect (&rest args)
>   (let ((format-alist nil)
> 	(auto-mode-alist (nnheader-auto-mode-alist))
> 	(default-major-mode 'fundamental-mode)
> 	(enable-local-variables nil)
>         (after-insert-file-functions nil))
>     (apply 'find-file-noselect args)))
> 
> Shouldn't this work under XEmacs as well?

Looks good to me.

-- 
Hrvoje Niksic <hniksic@srce.hr> | Student at FER Zagreb, Croatia
--------------------------------+--------------------------------
Oh lord won't you buy me a color TV...

Re: Possible security hole? nnfolder evaluates local variables..

[Lars Magne Ingebrigtsen]

Steven L Baur <steve@xemacs.org> writes:

> If this is the reason those functions are still there, I think it's
> safe to kill them now. :-)

Ok; I've killed the functions for Gnus v5.4.54.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen