Re: org-crypt bug and other org inconveniences.

[Thierry Volpiatto]

Stefan Monnier <address@hidden> writes:

>>> Note that this warning is not necessarily a problem.  Actually, I tend
>>> to think we should just get rid of this warning (which I added, a few
>>> years ago).
>> Yes, setting a var locally already let-bounded works actually.
>> But it is not what the documentation say.
>
>> ,----
>> | Making a variable buffer-local within a `let'-binding for that
>> | variable does not work reliably, unless the buffer in which you do
>> | this is not current either on entry to or exit from the `let'.
>> | This is because `let' does not distinguish between different kinds
>> | of bindings; it knows only which variable the binding was made for.
>> `----
>
> The documentation is indeed rather conservative.  But it is true that
> over the years many bugs were found in this area (and fixed).
> AFAIK there are no remaining bugs in this area (no proof of it, but
> I did try to consider "all possible cases"), *but* the "right behavior"
> in some cases is not always completely well-defined.  E.g.:
>
>     <...global value of x = 1...>
>     (let ((x 2))
>       (make-local-variable 'x))
>
> After this code (default-value 'x) should have value 1.
> But what about the buffer-local value of x?
>
> So the above comment basically tries to scare coders away from mixing
> let-bindings and buffer-local bindings, because "there be dragons".

Agree. I have fixed all my code last year according to your warning.
So probably it is good to keep the warning and fix all the code that
trigger it ?

The patch about org-crypt does this (It is in my .emacs for months now).

>>> BTW, why not try to push Helm's support for Org to Org, i.e. make Org
>>> support Helm, rather than the other way around?
>> Sure, but I don't think org developers want to maintain/develop code
>> depending on external applications.
>> Emacs developers too I think, (AFAIK sending a patch to emacs handling
>> external dependencies have been always refused).
>
> If the interface is sufficiently clean and generic, I'd definitely
> consider it.  Can't speak for Org people.
Ok I will remember it, thanks.

> PS: Last time we talked about including Helm in GNU ELPA, there were
> issues about copyright, IIRC.  If those issues are mostly in the various
> mode-specific support files, we could focus on integrating the
> infrastructure part of Helm separately.

I have modified now most of the code during last two years.
But part of this code is still from somebody else even if massively
modified, so I don't know.

-- 
Thierry
Get my Gnupg key:
gpg --keyserver pgp.mit.edu --recv-keys 59F29997