Re: add info about safe-local-variable to describe-variable

[Ken Manheimer]

On 5/12/06, Reiner Steib <address@hidden> wrote:
> On Fri, May 12 2006, Ken Manheimer wrote:
>
> > i've attached a patch (allout-patch.txt), ChangeLog
> > (allout-ChangeLog.txt), and revised NEWS fragment (allout-NEWS.txt)
> > for allout that addresses the stuff we've been discussing, and a few
> > other things.
>
> Thanks, I have committed your changes.

thanks!

> > in the patch, any autoloaded safe-local-variables lambdas are quoted.
> > the patch also incorporates some of the other refinements that were
> > suggested in this discussion, though i am leaving the
> > safe-local-variable property assignments as active code (rather than
> > containing them in the autoload comment)
>
> A drawback is that user's won't see the predicate in the output of
> `describe-variable':
>
> ,----
> | This variable is safe as a file local variable if its value
> | satisfies the predicate which is byte-compiled expression.
> `----

the lambdas are quoted, so they ought to show as text.  (they do for me.)

> > and continuing to use a lambda rather than booleanp.  both of these
> > measures will make it easier to use allout in older emacs, without
> > adding significant complexity.
>
> But here, I still don't think it makes much sense to provide backward
> compatibility for older CVS Emacs versions.  Not using `booleanp'
> makes the code and the output of `describe-variable' harder to read.

i don't understand.  older emacs 22 CVS versions are not my concern -
emacs 21 and earlier lack booleanp.

> If you can't update your CVS Emacs, you could define `booleanp' if
> it's not bound.

i debated providing a booleanp definition if not already fboundp, but
considered that more clutter than just using the lambda.  i can be
convinced otherwise, but it seems like a fairly close call, one way or
the other.
--
ken manheimer
address@hidden
http://myriadicity.net