at risk

at risk

[Andreas Röhler]

When execution might rely on variables set by user, resp. at a later time, such a function should preferably not be used by programs, i.e.
in circumstances being with some probability not aware of the changed proceeding.

This is the case for example with `forward-sexp', which might be taken by `forward-sexp-function'.

Please consider bug#14611. Being concerned a certain amount of code is set already at risk that way.

Thanks,

Andreas

Re: at risk

[Davis Herring]

> This is the case for example with `forward-sexp', which might be taken
> by `forward-sexp-function'.

...which is not a user option?  It's meant for a major mode to set
(precisely so that other code that uses `forward-sexp' works
transparently), not the user.

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: at risk

[Andreas Röhler]

Am 17.06.2013 17:45, schrieb Davis Herring:
>> This is the case for example with `forward-sexp', which might be taken
>> by `forward-sexp-function'.
>
> ...which is not a user option?  It's meant for a major mode to set
> (precisely so that other code that uses `forward-sexp' works
> transparently), not the user.
>
> Davis
>

If a major-mode sets this for some purpose - how to exclude the Emacs core, who uses this and should proceed the common way, not as set lately?

Andreas

Re: at risk

[Davis Herring]

> If a major-mode sets this for some purpose - how to exclude the Emacs
> core, who uses this and should proceed the common way, not as set lately?

Use `scan-sexps', which is the majority of the default implementation of
`forward-sexp', instead?

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: at risk

[Andreas Röhler]

Am 17.06.2013 21:45, schrieb Davis Herring:
>> If a major-mode sets this for some purpose - how to exclude the Emacs
>> core, who uses this and should proceed the common way, not as set lately?
>
> Use `scan-sexps', which is the majority of the default implementation of
> `forward-sexp', instead?#

Right, for example. For sure: forward-sexp is nicely designed as a command to tweak by users, that's a kind of stuff I like at Emacs.
Just avoid turning it into a bug-source, don't build in forward-sexp in Emacs Lisp core programs, as this answer indicates:

http://lists.gnu.org/archive/html/bug-gnu-emacs/2013-06/msg00338.html

Andreas
>
> Davis
>