turn off safe-local-variable-values checking

turn off safe-local-variable-values checking

[Shug Boabby]

As i'm sure you all know: the security procedures for local variables
were recently beefed up. Now, if a user ever wishes to store local
variables in a file, they must manually accept each variable on
opening (or store them in .emacs).

This actually really bugs me, and I'd really rather live with the
security risk... is there a way to turn this behaviour off by default
for all files? I cannot find any documentation on the new feature and
an Emacs Wiki search for the above mentioned variable gives me 1248
results.

Apologies if you think this is to the wrong list, my understanding is
that emacs.help is really only for officially released versions of
emacs.

Re: turn off safe-local-variable-values checking

[Richard Stallman]

    This actually really bugs me, and I'd really rather live with the
    security risk... is there a way to turn this behaviour off by default
    for all files? I cannot find any documentation on the new feature and
    an Emacs Wiki search for the above mentioned variable gives me 1248
    results.

We don't have an option to do this, and I am not sure we should add
one, but it should be pretty easy for you to edit the code to do what
you want.

Re: turn off safe-local-variable-values checking

[Stefan Monnier]

> As i'm sure you all know: the security procedures for local variables
> were recently beefed up.  Now, if a user ever wishes to store local
> variables in a file, they must manually accept each variable on
> opening (or store them in .emacs).

Right, they should not only add those vars's values in the corresponding
files, but also add them in their .emacs to safe-local-variable-values.
I don't see why it should be such a big deal.

Can you provide a scenario where it's really bothersome?


        Stefan

Re: turn off safe-local-variable-values checking

[Shug Boabby]

On 31/03/06, Stefan Monnier <monnier@iro.umontreal.ca> wrote:
> Can you provide a scenario where it's really bothersome?

oh... try opening up each section file of your thesis (each file
containing 3 or 4 AUCTeX variables), across several computers, and
getting that bloody box pop up and ask 4X questions!! (where X are the
number of files i have in desktop save)

Re: turn off safe-local-variable-values checking

[Shug Boabby]

On 31/03/06, Stefan Monnier <monnier@iro.umontreal.ca> wrote:
> Right, they should not only add those vars's values in the corresponding
> files, but also add them in their .emacs to safe-local-variable-values.
> I don't see why it should be such a big deal.

It actually really bothers me from an aesthetic point of view to have
such specific details in my ~/.emacs. I pride myself on my ~/.emacs
being hand-written and easy to edit... I don't like it when Emacs
automatically adds values itself. I'd rather learn about how those
variables effect my editing experience and add them myself after
careful consideration. Tell me you don't agree? :-)

If the opinion is that this behaviour is not to be disabled... may I
suggest that the allowed local variables be stored in another (user
definable) file in the same style as the ~/.abbrev_defs is not in the
~/.emacs? Or appended to .emacs.desktop... just keep it away from my
previous ~/.emacs!!!

Re: turn off safe-local-variable-values checking

[Dan Nicolaescu]

"Shug Boabby" <shug.boabby@gmail.com> writes:

  > On 31/03/06, Stefan Monnier <monnier@iro.umontreal.ca> wrote:
  > > Can you provide a scenario where it's really bothersome?
  > 
  > oh... try opening up each section file of your thesis (each file
  > containing 3 or 4 AUCTeX variables), across several computers, and
  > getting that bloody box pop up and ask 4X questions!! (where X are the
  > number of files i have in desktop save)

Try a newer versions of AUCTeX, the warning is disabled by default. I
am not sure what version started doing that, I use the CVS.

Yeah, it would be great if AUCTeX was integrated in emacs, so you
won't need to install it separately... 

Re: turn off safe-local-variable-values checking

[David Kastrup]

Dan Nicolaescu <dann@ics.uci.edu> writes:

> "Shug Boabby" <shug.boabby@gmail.com> writes:
>
>   > On 31/03/06, Stefan Monnier <monnier@iro.umontreal.ca> wrote:
>   > > Can you provide a scenario where it's really bothersome?
>   > 
>   > oh... try opening up each section file of your thesis (each file
>   > containing 3 or 4 AUCTeX variables), across several computers, and
>   > getting that bloody box pop up and ask 4X questions!! (where X are the
>   > number of files i have in desktop save)
>
> Try a newer versions of AUCTeX, the warning is disabled by
> default. I am not sure what version started doing that, I use the
> CVS.
>
> Yeah, it would be great if AUCTeX was integrated in emacs, so you
> won't need to install it separately...

It is still missing copyright assignments for that.  If somebody would
volunteer to invest more time than I currently manage for hunting
people down and asking them, this would be appreciated.  Don't hold
your breath for this to finish, but it is something that deserves
better progress than I have made so far.

Also I am not sure that the release schedule of Emacs and AUCTeX are
really compatible: not everybody using AUCTeX takes his Emacs fresh
from CVS.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: turn off safe-local-variable-values checking

[Dan Nicolaescu]

David Kastrup <dak@gnu.org> writes:

  > Dan Nicolaescu <dann@ics.uci.edu> writes:
  > 
  > > "Shug Boabby" <shug.boabby@gmail.com> writes:
  > >
  > >   > On 31/03/06, Stefan Monnier <monnier@iro.umontreal.ca> wrote:
  > >   > > Can you provide a scenario where it's really bothersome?
  > >   > 
  > >   > oh... try opening up each section file of your thesis (each file
  > >   > containing 3 or 4 AUCTeX variables), across several computers, and
  > >   > getting that bloody box pop up and ask 4X questions!! (where X are the
  > >   > number of files i have in desktop save)
  > >
  > > Try a newer versions of AUCTeX, the warning is disabled by
  > > default. I am not sure what version started doing that, I use the
  > > CVS.
  > >
  > > Yeah, it would be great if AUCTeX was integrated in emacs, so you
  > > won't need to install it separately...
  > 
  > It is still missing copyright assignments for that.  If somebody would
  > volunteer to invest more time than I currently manage for hunting
  > people down and asking them, this would be appreciated.  Don't hold
  > your breath for this to finish, but it is something that deserves
  > better progress than I have made so far.

Hopefully someone can help you with the copyrights, maybe posting a
message about this would help.
Can you say what the status is right now?
Is the core OK ? (meaning the latex.el, tex*.el files)
The style/* files are not that crucial, and they could be
reimplemented if needed.


  > Also I am not sure that the release schedule of Emacs and AUCTeX are
  > really compatible: not everybody using AUCTeX takes his Emacs fresh
  > from CVS.

True, but the main problem that I have seen is that a lot of emacs
users not even aware that AUCTeX exists...
WRT the release schedule, cc-mode and gnus seem to manage separate
releases and being integrated...

Re: turn off safe-local-variable-values checking

[David Kastrup]

Dan Nicolaescu <dann@ics.uci.edu> writes:

> David Kastrup <dak@gnu.org> writes:
>
>   > Dan Nicolaescu <dann@ics.uci.edu> writes:
>   > 
>   > > Yeah, it would be great if AUCTeX was integrated in emacs, so you
>   > > won't need to install it separately...
>   > 
>   > It is still missing copyright assignments for that.  If somebody would
>   > volunteer to invest more time than I currently manage for hunting
>   > people down and asking them, this would be appreciated.  Don't hold
>   > your breath for this to finish, but it is something that deserves
>   > better progress than I have made so far.
>
> Hopefully someone can help you with the copyrights, maybe posting a
> message about this would help.
> Can you say what the status is right now?

The three most fundamental contributors (who have been maintainers at
one time) have all agreed to a copyright assignment, but I've only
secured two: the third does not appear to respond to Email.

> Is the core OK ? (meaning the latex.el, tex*.el files)

Definitely not.  Even when the fundamental contributors will be
secured, there will be dozens of other people who should be rounded
up.  Most of their contributions could be replaced, however, should
the need arise.

> The style/* files are not that crucial, and they could be
> reimplemented if needed.

Possibly.  The newer style files have mostly been done by people with
assignments: we don't accept new contributions without one.

>   > Also I am not sure that the release schedule of Emacs and AUCTeX
>   > are really compatible: not everybody using AUCTeX takes his
>   > Emacs fresh from CVS.
>
> True, but the main problem that I have seen is that a lot of emacs
> users not even aware that AUCTeX exists...

I think the coverage among readers of comp.text.tex is pretty good.
We are also making it reasonably easy for system administrators to
have AUCTeX active by default on their, while individual users can
easily customize it off again for some or all TeX-related modes if
they really want to.  This also helps: most people accept the site
defaults, and having an AUCTeX that can be turned off again (for the
sake of Emacs' default modes) makes activation by default a reasonable
site policy.

> WRT the release schedule, cc-mode and gnus seem to manage separate
> releases and being integrated...

Somewhat.  But it leads to people using old gnus.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: turn off safe-local-variable-values checking

[Stefan Monnier]

>> Can you provide a scenario where it's really bothersome?

> oh... try opening up each section file of your thesis (each file
> containing 3 or 4 AUCTeX variables), across several computers, and
> getting that bloody box pop up and ask 4X questions!! (where X are the
> number of files i have in desktop save)

How many different variable/value pairs are we talking about?
The assumption is that you're going to add those to your
safe-local-variable-values list in your .emacs so that the question won't be
asked.
Or if the variables are safe, they should simply be marked as such by the
package's author.


        Stefan

Re: turn off safe-local-variable-values checking

[Johan Bockgård]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> The assumption is that you're going to add those to your
> safe-local-variable-values list in your .emacs so that the question
> won't be asked. Or if the variables are safe, they should simply be
> marked as such by the package's author.

Emacs warns about some files that are included in Emacs itself, namely

table.el
allout.el
ediff-diff.el
ediff-init.el
ediff-merg.el
ediff-mult.el
ediff-ptch.el
ediff-util.el
ediff-vers.el
ediff-wind.el
ediff.el

-- 
Johan Bockgård

Re: turn off safe-local-variable-values checking

[Richard Stallman]

    Emacs warns about some files that are included in Emacs itself, namely

    table.el
    allout.el
    ediff*.el

This is useful--I presume these bindings are safe,
so we should define predicates for the variables they bind.

Would someone like to do that now, and ack?

Re: turn off safe-local-variable-values checking

[Andreas Schwab]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

> Or if the variables are safe, they should simply be marked as such by the
> package's author.

I have the following settings in my .emacs:

(put 'left-margin 'safe-local-variable t)
(put 'byte-compile-dynamic 'safe-local-variable t)
(put 'byte-compile-warnings 'safe-local-variable t)

I think they should generally be safe for all values and marked as such.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
PGP key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: turn off safe-local-variable-values checking

[Richard Stallman]

    (put 'left-margin 'safe-local-variable t)
    (put 'byte-compile-dynamic 'safe-local-variable t)
    (put 'byte-compile-warnings 'safe-local-variable t)

Would you please set them up as safe?

AUCTex (was: turn off safe-local-variable-values checking)

[Dan Nicolaescu]

David Kastrup <dak@gnu.org> writes:

  > Dan Nicolaescu <dann@ics.uci.edu> writes:
  > 
  > > David Kastrup <dak@gnu.org> writes:
  > >
  > >   > Dan Nicolaescu <dann@ics.uci.edu> writes:

  > > True, but the main problem that I have seen is that a lot of emacs
  > > users not even aware that AUCTeX exists...
  > 
  > I think the coverage among readers of comp.text.tex is pretty good.

That might be true, but readers of comp.text.tex are a small minority
(and probably the most technically astute one) among  TeX users. Well,
sadly newsgroup readers are a small minority in general nowadays. So I
think my original point still stands. 

  > We are also making it reasonably easy for system administrators to
  > have AUCTeX active by default on their, while individual users can
  > easily customize it off again for some or all TeX-related modes if
  > they really want to.  This also helps: most people accept the site
  > defaults, and having an AUCTeX that can be turned off again (for the
  > sake of Emacs' default modes) makes activation by default a reasonable
  > site policy.

That is nice. Unfortunately of the 6 sites that I personally know
about, none has AUCTeX installed :-(

Anyway, the main question is: Do AUCTeX maintainers want to integrate
it in Emacs anytime soon? (Or at all?)

IMHO having AUCTeX integrated would be beneficial to both AUCTeX as it
would reach a wider audience, and also to Emacs as it would help users
better manage their tex documents.

I am sure that if there's a strong desire to integrate AUCTeX, you
will get some help.

Thanks

        --dan

Re: turn off safe-local-variable-values checking

[Andreas Schwab]

Richard Stallman <rms@gnu.org> writes:

>     (put 'left-margin 'safe-local-variable t)
>     (put 'byte-compile-dynamic 'safe-local-variable t)
>     (put 'byte-compile-warnings 'safe-local-variable t)
>
> Would you please set them up as safe?

Done.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
PGP key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: turn off safe-local-variable-values checking

[Reiner Steib]

On Tue, Apr 04 2006, Richard Stallman wrote:

>     (put 'left-margin 'safe-local-variable t)
>     (put 'byte-compile-dynamic 'safe-local-variable t)
>     (put 'byte-compile-warnings 'safe-local-variable t)
>
> Would you please set them up as safe?

--- files.el	4 Apr 2006 18:58:48 -0000	1.821
+++ files.el	4 Apr 2006 19:28:08 -0000
@@ -2333,12 +2333,14 @@
   (eval
    `(mapc (lambda (pair)
 	    (put (car pair) 'safe-local-variable (cdr pair)))
-	  '((byte-compile-dynamic . t)
+	  '((abbrev-mode        .  t)
+	    (byte-compile-dynamic . t)
 	    (byte-compile-dynamic-docstrings . t)
 	    (byte-compile-warnings . t)
 	    (c-basic-offset     .  integerp)
 	    (c-file-style       .  stringp)
 	    (c-indent-level     .  integerp)
+	    (colon-double-space .  t)
 	    (comment-column     .  integerp)
 	    (compile-command    . ,string-or-null)
 	    (fill-column        .  integerp)
@@ -2355,6 +2357,11 @@
 	    (page-delimiter     . ,string-or-null)
 	    (paragraph-start    . ,string-or-null)
 	    (paragraph-separate . ,string-or-null)
+	    ;; textmode/reftex-vars.el
+	    (reftex-fref-is-default  . t)
+	    (reftex-guess-label-type . t)
+	    (reftex-level-indent     . integerp)
+	    (reftex-vref-is-default  . t)
 	    (sentence-end       . ,string-or-null)
 	    (sentence-end-double-space . t)
 	    (tab-width          .  integerp)

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: turn off safe-local-variable-values checking

[Stefan Monnier]

> +	    ;; textmode/reftex-vars.el
> +	    (reftex-fref-is-default  . t)
> +	    (reftex-guess-label-type . t)
> +	    (reftex-level-indent     . integerp)
> +	    (reftex-vref-is-default  . t)

I think these settings belong in reftex*.el rather than in files.el.


        Stefan

Re: AUCTex

[David Kastrup]

Dan Nicolaescu <dann@ics.uci.edu> writes:

> Anyway, the main question is: Do AUCTeX maintainers want to
> integrate it in Emacs anytime soon? (Or at all?)

I answered that question already in the previous post, but you snipped
the reply.  AUCTeX is not going to be integrated into Emacs anytime
soon due to outstanding copyright assignments, and the process of
collecting them is progressing rather slowly.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: turn off safe-local-variable-values checking

[Richard Stallman]

Does anyone else want to suggest some variables we should mark as safe?

Re: turn off safe-local-variable-values checking

[Miles Bader]

Richard Stallman <rms@gnu.org> writes:
> Does anyone else want to suggest some variables we should mark as safe?

paragraph-ignore-fill-prefix

-miles
-- 
Somebody has to do something, and it's just incredibly pathetic that it
has to be us.  -- Jerry Garcia

Re: turn off safe-local-variable-values checking

[Reiner Steib]

On Tue, Apr 04 2006, Stefan Monnier wrote:

>> +	    ;; textmode/reftex-vars.el
>> +	    (reftex-fref-is-default  . t)
>> +	    (reftex-guess-label-type . t)
>> +	    (reftex-level-indent     . integerp)
>> +	    (reftex-vref-is-default  . t)
>
> I think these settings belong in reftex*.el rather than in files.el.

The comment in `files.el' says it should be set there for variables
defined by minor modes and `reftex-mode' is a minor mode:

;; For variables defined by minor modes, put the safety declarations
;; here, not in the file defining the minor mode (when Emacs visits a
;; file specifying that local variable, the minor mode file may not be
;; loaded yet).  For variables defined by major modes, the safety
;; declarations can go into the major mode's file, since that will be
;; loaded before file variables are processed.

CMIIW, please.  And maybe clarify the comment if I was missing
something.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: turn off safe-local-variable-values checking

[Stefan Monnier]

> ;; For variables defined by minor modes, put the safety declarations
> ;; here, not in the file defining the minor mode (when Emacs visits a
> ;; file specifying that local variable, the minor mode file may not be
> ;; loaded yet).  For variables defined by major modes, the safety
> ;; declarations can go into the major mode's file, since that will be
> ;; loaded before file variables are processed.

Interesting.  I'd have to disagree: put it/them in the minor mode's file,
but with an ;;;###autoload cookie.

I think it's important to put those safe-local-variable properties at the
same spot as where the variable is declared.


        Stefan

Re: turn off safe-local-variable-values checking

[Richard Stallman]

    > Does anyone else want to suggest some variables we should mark as safe?

    paragraph-ignore-fill-prefix

Please mark it.

Someone made the point that it would be better to put these markings
next to the definitions of the variables.  Could you do it that way?

Re: turn off safe-local-variable-values checking

[Chong Yidong]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> ;; For variables defined by minor modes, put the safety declarations
>> ;; here, not in the file defining the minor mode (when Emacs visits a
>> ;; file specifying that local variable, the minor mode file may not be
>> ;; loaded yet).  For variables defined by major modes, the safety
>> ;; declarations can go into the major mode's file, since that will be
>> ;; loaded before file variables are processed.
>
> Interesting.  I'd have to disagree: put it/them in the minor mode's file,
> but with an ;;;###autoload cookie.

Yes, good idea.

Re: AUCTex

[Dan Nicolaescu]

David Kastrup <dak@gnu.org> writes:

  > Dan Nicolaescu <dann@ics.uci.edu> writes:
  > 
  > > Anyway, the main question is: Do AUCTeX maintainers want to
  > > integrate it in Emacs anytime soon? (Or at all?)
  > 
  > I answered that question already in the previous post, but you snipped
  > the reply.  AUCTeX is not going to be integrated into Emacs anytime
  > soon due to outstanding copyright assignments, and the process of
  > collecting them is progressing rather slowly.

So what is the way forward for this? 
What can be done to improve the situation? 

Would distributing AUCTeX like leim was distributed in previous emacs
versions help? (i.e. leim distributed separately, but it could be
unpacked in the emacs source tree and compiled and installed
together). Just an idea...

Re: AUCTex

[David Kastrup]

Dan Nicolaescu <dann@ics.uci.edu> writes:

> David Kastrup <dak@gnu.org> writes:
>
>   > Dan Nicolaescu <dann@ics.uci.edu> writes:
>   > 
>   > > Anyway, the main question is: Do AUCTeX maintainers want to
>   > > integrate it in Emacs anytime soon? (Or at all?)
>   > 
>   > I answered that question already in the previous post, but you snipped
>   > the reply.  AUCTeX is not going to be integrated into Emacs anytime
>   > soon due to outstanding copyright assignments, and the process of
>   > collecting them is progressing rather slowly.
>
> So what is the way forward for this? 
> What can be done to improve the situation?

Basically collecting assignments.

> Would distributing AUCTeX like leim was distributed in previous
> emacs versions help? (i.e. leim distributed separately, but it could
> be unpacked in the emacs source tree and compiled and installed
> together). Just an idea...

AUCTeX is a separate project and works with a variety of Emacsen.  It
is also released more often than every 5 years or so.  I don't think
that at the current point of time there is much room for improvement.
It might be possible to create a tarball for dropping into an Emacs
tree before compiling and autoload generation (which is somewhat
similar to what we do to create an XEmacs package).  But I don't see
it change the situation much.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: turn off safe-local-variable-values checking

[Richard Stallman]

    > ;; For variables defined by minor modes, put the safety declarations
    > ;; here, not in the file defining the minor mode (when Emacs visits a
    > ;; file specifying that local variable, the minor mode file may not be
    > ;; loaded yet).  For variables defined by major modes, the safety
    > ;; declarations can go into the major mode's file, since that will be
    > ;; loaded before file variables are processed.

    Interesting.  I'd have to disagree: put it/them in the minor mode's file,
    but with an ;;;###autoload cookie.

That seems like a good method to achieve the goal
(which is that the property should always be there).
Would you please edit the comment?

    I think it's important to put those safe-local-variable properties at the
    same spot as where the variable is declared.

I agree, that will be better for maintenance, in case a change in the meaning
of the variable calls for a change in the property.

Additional safe-local-variables (was: turn off safe-local-variable-values checking)

[Reiner Steib]

On Thu, Apr 06 2006, Richard Stallman wrote:

>     Interesting.  I'd have to disagree: put it/them in the minor mode's file,
>     but with an ;;;###autoload cookie.
>
> That seems like a good method to achieve the goal
> (which is that the property should always be there).
> Would you please edit the comment?
>
>     I think it's important to put those safe-local-variable properties at the
>     same spot as where the variable is declared.
>
> I agree, that will be better for maintenance, in case a change in the meaning
> of the variable calls for a change in the property.

I've done that for the new variable I've suggested and for some other
variable (I didn't have time to process all of the list), e.g.:

;;;###autoload(put 'paragraph-separate 'safe-local-variable 'stringp)

`page-delimiter' and `paragraph-start' were marked with string-or-null
whereas the custom type doesn't allow nil.  I used stringp then.  Feel
free to adjust the autoloads or the defcustoms.

Maybe it would make sense to introduce `string-or-null-p' to avoid
using (lambda (a) (or (stringp a) (null a))) in the autoloads.

BTW, wouldn't it be possible to integrate the safe-local-variable
properties into defcustom?

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: Additional safe-local-variables (was: turn off safe-local-variable-values checking)

[Richard Stallman]

    Maybe it would make sense to introduce `string-or-null-p' to avoid
    using (lambda (a) (or (stringp a) (null a))) in the autoloads.

Please do.

    BTW, wouldn't it be possible to integrate the safe-local-variable
    properties into defcustom?

Yes, I think we should.  We can have a keyword :safe whose value can
be either t (all values permitted by the custom type are safe) or a
predicate.  When it is t, I think Emacs should really check that the
local value fits the custom type.

However, one inconvenience in using this is that safety specified
in this way will only work once the defcustom is loaded.  For minor modes,
we will still need to do it differently (such as with a `put' call
in an autoload cookie).

Re: Additional safe-local-variables

[Reiner Steib]

On Thu, Apr 06 2006, Richard Stallman wrote:

>     Maybe it would make sense to introduce `string-or-null-p' to avoid
>     using (lambda (a) (or (stringp a) (null a))) in the autoloads.
>
> Please do.

I added it to `subr.el' in the "Misc. useful functions." part:

(defun string-or-null-p (object)
  "Return t if OBJECT is a string or nil.
Otherwise, return nil."
  (or (stringp object) (null object)))

(Should it be `defun' or `defsubst'?)

I also updated the comment about safe-local-variable declarations in
`files.el'.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: Additional safe-local-variables

[Richard Stallman]

    I added it to `subr.el' in the "Misc. useful functions." part:

    (defun string-or-null-p (object)
      "Return t if OBJECT is a string or nil.
    Otherwise, return nil."
      (or (stringp object) (null object)))

defun is best here, I think; we are not concerned about speed.

Could you please add this to NEWS and the Lisp Manual?

Re: Additional safe-local-variables

[Reiner Steib]

On Fri, Apr 07 2006, Richard Stallman wrote:

>     I added it to `subr.el' in the "Misc. useful functions." part:
>
>     (defun string-or-null-p (object)
>       "Return t if OBJECT is a string or nil.
>     Otherwise, return nil."
>       (or (stringp object) (null object)))
>
> defun is best here, I think; we are not concerned about speed.
>
> Could you please add this to NEWS and the Lisp Manual?

I've added it under "General Lisp changes" in NEWS and in strings.texi
(Predicates for Strings).  I hope this makes sense.

Bye, Reiner.
-- 
       ,,,
      (o o)
---ooO-(_)-Ooo---  |  PGP key available  |  http://rsteib.home.pages.de/

Re: Additional safe-local-variables

[Richard Stallman]

    I've added it under "General Lisp changes" in NEWS and in strings.texi
    (Predicates for Strings).  I hope this makes sense.

Yes, thanks.

Re: turn off safe-local-variable-values checking

[Richard Stallman]

[I sent this message a week ago but did not get a response.]

    Emacs warns about some files that are included in Emacs itself, namely

    table.el
    allout.el
    ediff*.el

This is useful--I presume these bindings are safe,
so we should define predicates for the variables they bind.

Would someone like to do that now, and ack?

Re: turn off safe-local-variable-values checking

[Stuart D. Herring]

>     Emacs warns about some files that are included in Emacs itself, namely
>
>     table.el
>     allout.el
>     ediff*.el
>
> This is useful--I presume these bindings are safe,
> so we should define predicates for the variables they bind.
>
> Would someone like to do that now, and ack?

I'll do this.  But a question, first: the docs for
`enable-local-variables' says (since 1998) that `normal-mode' called
interactively always obeys local variable specifications, when it in fact
only acts as if `enable-local-variables' is t, which obeys only safe
variables automatically.  Should the doc be updated?

Davis

-- 
This product is sold by volume, not by mass.  If it appears too dense or
too sparse, it is because mass-energy conversion has occurred during
shipping.

Re: turn off safe-local-variable-values checking

[Stuart D. Herring]

[-- Attachment #1: Type: text/plain, Size: 805 bytes --]

>     table.el
>     allout.el
>     ediff*.el
>
> This is useful--I presume these bindings are safe,
> so we should define predicates for the variables they bind.
>
> Would someone like to do that now, and ack?

Interesting: each of those three files required a different approach.  I
added predicates for time-stamp.el's local variables, putting reasonable
limits on the integers.  Allout just needed to autoload its
already-present safe-local declaration.  And ediff was using part of
edebug which could really be used for evil, so I wrote a check for
non-evil use and added that to the code for symbol properties.

Patch attached, natch.

Davis

-- 
This product is sold by volume, not by mass.  If it appears too dense or
too sparse, it is because mass-energy conversion has occurred during
shipping.

[-- Attachment #2: local-variables.patch --]
[-- Type: application/octet-stream, Size: 6711 bytes --]

? local-variables.patch
Index: lisp/ChangeLog
===================================================================
RCS file: /sources/emacs/emacs/lisp/ChangeLog,v
retrieving revision 1.9390
diff -a -c -r1.9390 ChangeLog
*** lisp/ChangeLog	12 Apr 2006 05:11:12 -0000	1.9390
--- lisp/ChangeLog	12 Apr 2006 19:55:17 -0000
***************
*** 1,3 ****
--- 1,18 ----
+ 2006-04-12  Davis Herring  <herring@lanl.gov>
+ 
+ 	* files.el: Add predicates for time-stamp.el's file-local
+ 	variables (here, since it's never automatically loaded).  Remove
+ 	lingering `string-or-null' let-binding.
+ 	(hack-one-local-variable-eval-safep): Recognize
+ 	`edebug-form-spec' for `put', but only if it passes
+ 	`edebug-basic-spec'.  Generalize `put' handling.
+ 
+ 	* emacs-lisp/edebug.el (edebug-basic-spec): New function for
+ 	vetting file-local form specs.
+ 
+ 	* allout.el (allout-layout): Autoload its `safe-local-variable'
+ 	property.
+ 
  2006-04-12  Stefan Monnier  <monnier@iro.umontreal.ca>
  
  	* progmodes/perl-mode.el (perl-indent-new-calculate):
Index: lisp/allout.el
===================================================================
RCS file: /sources/emacs/emacs/lisp/allout.el,v
retrieving revision 1.70
diff -a -c -r1.70 allout.el
*** lisp/allout.el	15 Mar 2006 03:14:15 -0000	1.70
--- lisp/allout.el	12 Apr 2006 19:55:18 -0000
***************
*** 666,671 ****
--- 666,672 ----
  `allout-layout' can additionally have the value `t', in which
  case the value of `allout-default-layout' is used.")
  (make-variable-buffer-local 'allout-layout)
+ ;;;###autoload
  (put 'allout-layout 'safe-local-variable t)
  
  ;;;_  : Topic header format
Index: lisp/files.el
===================================================================
RCS file: /sources/emacs/emacs/lisp/files.el,v
retrieving revision 1.825
diff -a -c -r1.825 files.el
*** lisp/files.el	11 Apr 2006 17:57:34 -0000	1.825
--- lisp/files.el	12 Apr 2006 19:55:18 -0000
***************
*** 2343,2371 ****
  ;; For variables defined in the C source code the declaration should go here:
  
  ;; FIXME: Some variables should be moved according to the rules above.
! (let ((string-or-null (lambda (a) (or (stringp a) (null a)))))
!   (eval
!    `(mapc (lambda (pair)
! 	    (put (car pair) 'safe-local-variable (cdr pair)))
! 	  '((byte-compile-dynamic . t)
! 	    (byte-compile-dynamic-docstrings . t)
! 	    (byte-compile-warnings . t)
! 	    (c-basic-offset     .  integerp)
! 	    (c-file-style       .  stringp)
! 	    (c-indent-level     .  integerp)
! 	    (comment-column     .  integerp)
! 	    (compile-command    .  string-or-null-p)
! 	    (fill-column        .  integerp)
! 	    (fill-prefix        .  string-or-null-p)
! 	    (indent-tabs-mode   .  t)
! 	    (kept-new-versions  .  integerp)
! 	    (left-margin        .  t)
! 	    (no-byte-compile    .  t)
! 	    (no-update-autoloads . t)
! 	    (outline-regexp     .  string-or-null-p)
! 	    (tab-width          .  integerp) ;; C source code
! 	    (truncate-lines     .  t) ;; C source code
! 	    (version-control    .  t)))))
  
  (put 'c-set-style 'safe-local-eval-function t)
  
--- 2343,2380 ----
  ;; For variables defined in the C source code the declaration should go here:
  
  ;; FIXME: Some variables should be moved according to the rules above.
! (eval
!  `(mapc (lambda (pair)
! 	  (put (car pair) 'safe-local-variable (cdr pair)))
! 	'((byte-compile-dynamic . t)
! 	  (byte-compile-dynamic-docstrings . t)
! 	  (byte-compile-warnings . t)
! 	  (c-basic-offset     .  integerp)
! 	  (c-file-style       .  stringp)
! 	  (c-indent-level     .  integerp)
! 	  (comment-column     .  integerp)
! 	  (compile-command    .  string-or-null-p)
! 	  (fill-column        .  integerp)
! 	  (fill-prefix        .  string-or-null-p)
! 	  (indent-tabs-mode   .  t)
! 	  (kept-new-versions  .  integerp)
! 	  (left-margin        .  t)
! 	  (no-byte-compile    .  t)
! 	  (no-update-autoloads . t)
! 	  (outline-regexp     .  string-or-null-p)
! 	  (tab-width          .  integerp) ;; C source code
! 	  (time-stamp-inserts-lines . t)
! 	  (time-stamp-line-limit . (lambda (val) (and (integerp val)
! 						      (not (zerop val))
! 						      (<= (abs val) 1000))))
! 	  (time-stamp-count   .  (lambda (val) (and (integerp val)
! 						    (<= val 100))))
! 	  (time-stamp-end     .  stringp)
! 	  (time-stamp-format  .  stringp)
! 	  (time-stamp-pattern .  string-or-null-p)
! 	  (time-stamp-start   .  stringp)
! 	  (truncate-lines     .  t) ;; C source code
! 	  (version-control    .  t))))
  
  (put 'c-set-style 'safe-local-eval-function t)
  
***************
*** 2689,2700 ****
        (and (eq (car exp) 'put)
  	   (hack-one-local-variable-quotep (nth 1 exp))
  	   (hack-one-local-variable-quotep (nth 2 exp))
! 	   (memq (nth 1 (nth 2 exp))
! 		 '(lisp-indent-hook))
! 	   ;; Only allow safe values of lisp-indent-hook;
! 	   ;; not functions.
! 	   (or (numberp (nth 3 exp))
! 	       (equal (nth 3 exp) ''defun)))
        ;; Allow expressions that the user requested.
        (member exp safe-local-eval-forms)
        ;; Certain functions can be allowed with safe arguments
--- 2698,2711 ----
        (and (eq (car exp) 'put)
  	   (hack-one-local-variable-quotep (nth 1 exp))
  	   (hack-one-local-variable-quotep (nth 2 exp))
! 	   (let ((prop (nth 1 (nth 2 exp))) (val (nth 3 exp)))
! 	     (cond ((eq prop 'lisp-indent-hook)
! 		    ;; Only allow safe values of lisp-indent-hook;
! 		    ;; not functions.
! 		    (or (numberp val) (equal val ''defun)))
! 		   ((eq prop 'edebug-form-spec)
! 		    ;; Only allow indirect form specs.
! 		    (edebug-basic-spec val)))))
        ;; Allow expressions that the user requested.
        (member exp safe-local-eval-forms)
        ;; Certain functions can be allowed with safe arguments
Index: lisp/emacs-lisp/edebug.el
===================================================================
RCS file: /sources/emacs/emacs/lisp/emacs-lisp/edebug.el,v
retrieving revision 3.88
diff -a -c -r3.88 edebug.el
*** lisp/emacs-lisp/edebug.el	3 Mar 2006 11:55:43 -0000	3.88
--- lisp/emacs-lisp/edebug.el	12 Apr 2006 19:55:19 -0000
***************
*** 258,263 ****
--- 258,277 ----
      edebug-form-spec
      ))
  
+ ;;;###autoload
+ (defun edebug-basic-spec (spec)
+   "Return t if SPEC uses only extant spec symbols.
+ An extant spec symbol is a symbol that is not a function and has a
+ `edebug-form-spec' property."
+   (cond ((listp spec)
+ 	 (catch 'basic
+ 	   (while spec
+ 	     (unless (edebug-basic-spec (car spec)) (throw 'basic nil))
+ 	     (setq spec (cdr spec)))
+ 	   t))
+ 	((symbolp spec)
+ 	 (unless (functionp spec) (get spec 'edebug-form-spec)))))
+ 
  ;;; Utilities
  
  ;; Define edebug-gensym - from old cl.el

[-- Attachment #3: Type: text/plain, Size: 142 bytes --]

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/emacs-devel

Re: turn off safe-local-variable-values checking

[Richard Stallman]

    I'll do this.  But a question, first: the docs for
    `enable-local-variables' says (since 1998) that `normal-mode' called
    interactively always obeys local variable specifications, when it in fact
    only acts as if `enable-local-variables' is t, which obeys only safe
    variables automatically.  Should the doc be updated?

I think normal-mode, when called manually, should always obey all the
file local variables.  But what do others think?

Re: turn off safe-local-variable-values checking

[Stuart D. Herring]

> I think normal-mode, when called manually, should always obey all the
> file local variables.  But what do others think?

I've been opposed to automatic file locals in the past, so I may be
biased.  But my thoughts are that normal-mode should do exactly what
find-file does automatically (so you can use it if you change the local
variables, or add a #! line, etc., without security concerns), and then if
we really want to have a "force everything", use C-u M-x normal-mode. 
(normal-mode does not currently use a prefix argument.)  Another (minor)
reason it's useful to call normal-mode without forcing all variables is
that it sets fundamental-mode if the current buffer just inherited a mode
(via default-major-mode being nil).

Davis

-- 
This product is sold by volume, not by mass.  If it appears too dense or
too sparse, it is because mass-energy conversion has occurred during
shipping.

M-x normal-mode and local variables

[Richard Stallman]

I'm interested in hearing more opinions about whether M-x normal-mode
should query, or just install all local variables.

Re: M-x normal-mode and local variables

[Stefan Monnier]

> I'm interested in hearing more opinions about whether M-x normal-mode
> should query, or just install all local variables.

I think using M-x normal-mode is sufficiently uncommon that we shouldn't try
too hard to streamline it.  So I vote to have it query.  Especially since
there's no guarantee that the user actually checked the local variables
before calling M-x normal-mode, so blindly obeying the local variables could
be dangerous in some scenarios.

After all, if the query is annoying, the answer is not to make M-x
normal-mode less interactive but to add the relevant var/val pairs to
safe-local-variable-values.


        Stefan

Re: M-x normal-mode and local variables

[Richard Stallman]

    I think using M-x normal-mode is sufficiently uncommon that we shouldn't try
    too hard to streamline it.  So I vote to have it query.  Especially since
    there's no guarantee that the user actually checked the local variables
    before calling M-x normal-mode, so blindly obeying the local variables could
    be dangerous in some scenarios.

That is a significant point.  I would still like to hear what others
have to say.

Re: turn off safe-local-variable-values checking

[Stuart D. Herring]

[-- Attachment #1: Type: text/plain, Size: 1686 bytes --]

[Original thread is at
http://lists.gnu.org/archive/html/emacs-devel/2006-04/msg00521.html.]

I wrote, a long time ago:

>>     table.el
>>     allout.el
>>     ediff*.el
>>
>> This is useful--I presume these bindings are safe,
>> so we should define predicates for the variables they bind.
>>
>> Would someone like to do that now, and ack?
>
> Interesting: each of those three files required a different approach.  I
> added predicates for time-stamp.el's local variables, putting reasonable
> limits on the integers.  Allout just needed to autoload its
> already-present safe-local declaration.  And ediff was using part of
> edebug which could really be used for evil, so I wrote a check for
> non-evil use and added that to the code for symbol properties.
>
> Patch attached, natch.

Oddly, most but not all of my patch was applied.  The part that wasn't
later became horribly out of date because of conflicting changes, but I've
resolved the conflicts several times and kept it alive, and am
resubmitting just the part that was missed as a diff against current
sources.  See the ChangeLog entry:

 	* files.el: Add predicates for time-stamp.el's file-local
 	variables (here, since it's never automatically loaded).

Assuming that there's nothing wrong with adding the rest of what was added
then, please install?

Davis

PS - Sadly, I was reminded of these changes by the recent thread
http://lists.gnu.org/archive/html/emacs-devel/2007-02/msg01042.html: there
was an insidious bug in the part of my patch that was applied!

-- 
This product is sold by volume, not by mass.  If it appears too dense or
too sparse, it is because mass-energy conversion has occurred during
shipping.

[-- Attachment #2: local-variables-2.patch --]
[-- Type: application/octet-stream, Size: 1105 bytes --]

Index: files.el
===================================================================
RCS file: /sources/emacs/emacs/lisp/files.el,v
retrieving revision 1.885
diff -c -r1.885 files.el
*** files.el	25 Feb 2007 09:24:11 -0000	1.885
--- files.el	1 Mar 2007 00:20:43 -0000
***************
*** 2449,2454 ****
--- 2448,2464 ----
  	(left-margin                     . integerp) ;; C source code
  	(no-update-autoloads             . booleanp)
  	(tab-width                       . integerp) ;; C source code
+ 	(time-stamp-inserts-lines        . t)
+ 	(time-stamp-line-limit           . (lambda (val)
+ 					     (and (integerp val)
+ 						  (not (zerop val))
+ 						  (<= (abs val) 1000))))
+ 	(time-stamp-count                . (lambda (val) (and (integerp val)
+ 							       (<= val 100))))
+ 	(time-stamp-end                  . stringp)
+ 	(time-stamp-format               . stringp)
+ 	(time-stamp-pattern              . string-or-null-p)
+ 	(time-stamp-start                . stringp)
  	(truncate-lines                  . booleanp))) ;; C source code
  
  (put 'c-set-style 'safe-local-eval-function t)

[-- Attachment #3: Type: text/plain, Size: 142 bytes --]

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/emacs-devel

Re: turn off safe-local-variable-values checking

[Chong Yidong]

"Stuart D. Herring" <herring@lanl.gov> writes:

> Oddly, most but not all of my patch was applied.  The part that wasn't
> later became horribly out of date because of conflicting changes, but I've
> resolved the conflicts several times and kept it alive, and am
> resubmitting just the part that was missed as a diff against current
> sources.  See the ChangeLog entry:
>
>  	* files.el: Add predicates for time-stamp.el's file-local
>  	variables (here, since it's never automatically loaded).

The safe-local-variable declarations were moved to time-stamp.el as
autoload forms.  See the comment in files.el.

So this patch is no longer necessary.

Re: turn off safe-local-variable-values checking

[Stuart D. Herring]

>> Oddly, most but not all of my patch was applied.  The part that wasn't
>> later became horribly out of date because of conflicting changes, but
>> I've
>> resolved the conflicts several times and kept it alive, and am
>> resubmitting just the part that was missed as a diff against current
>> sources.  See the ChangeLog entry:
>>
>>  	* files.el: Add predicates for time-stamp.el's file-local
>>  	variables (here, since it's never automatically loaded).
>
> The safe-local-variable declarations were moved to time-stamp.el as
> autoload forms.  See the comment in files.el.
>
> So this patch is no longer necessary.

Ah, thanks.  As it happens, I didn't use quite the same predicates as RMS
did when he added those, but I don't think anything worse than an
interruptible forever-long regexp search can result from those variables,
so I guess it's not of any great importance.

Davis

-- 
This product is sold by volume, not by mass.  If it appears too dense or
too sparse, it is because mass-energy conversion has occurred during
shipping.