Re: Dynamic modules: MODULE_HANDLE_SIGNALS etc.

[Daniel Colascione <dancol@dancol.org>]

[-- Attachment #1: Type: text/plain, Size: 872 bytes --]

On 12/21/2015 08:48 PM, Paul Eggert wrote:
> Daniel Colascione wrote:
>> I'd rather
>> Emacs just die on C stack overflow
>
> That would certainly be easier to implement! But as a user I would not
> find it acceptable.

You find it acceptable in most other programs. What problem are we
trying to solve with this overflow checking? We should do something
about unbound recursion in Lisp, but in C? Why? We expect a much higher
level of robustness in the C code.

Besides, we already crash if we overflow the stack while we're GCing.

>>  except when we know we're running
>> Lisp in such a way that we know we can recover.
>
> I'm afraid that's not good enough, as stack overflow can occur while
> running C code.

Yes, and the proper response to stack overflow in C code should be to
crash. We shouldn't be allocating stack without bound in C.



[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]