Re: Dynamic modules: emacs-module.c and signaling errors

[Stefan Monnier]

>> The general design is good, but forcing this signal-handling is just
>> brain-dead from where I stand.  I stated this many times.  And I'm
>> shocked that we'd follow Daniel's design on this detail just because he
>> advocates it.
> So we should follow your design because _you_ advocate it?

Well, yes.  All things being equal (i.e. just hot air on either side),
I'd assume that years-of-service and megabytes-of-contributions should
tilt the balance in my favor.

> And it's not "just" because I advocate it --- there's a good reason that
> the vast majority of general-purpose C libraries don't use non-local
> control flow for general-purpose error reporting.

Emacs is not a general purpose C library, so I don't see why that'd
be relevant.


        Stefan