Re: Lisp watchpoints

[Noam Postavsky]

On Sun, Nov 29, 2015 at 6:28 AM, Andreas Schwab <address@hidden> wrote:
>> Isn't the FUNCTIONP test needed for basically the same reason as using
>> CONSP over !NILP? i.e. to avoid segfault in case of (put <var>
>> 'watchers '([1])).
>
> Ffuncall does all necessary checks.

It would signal in case of error right? Currently the code is ignoring
bad watcher elements (including out of range integers), but perhaps
it's actually better to signal an error.