Re: [PATCH updated] Support for filesystem watching (inotify)

[Stefan Monnier <monnier@iro.umontreal.ca>]

>> My preference would be to wait for 24.2 since we're in feature freeze,
>> but it's the kind of feature which benefits from being available early
>> so other packages can start using it.  WDYT?
> I briefly looked at it for integer-overflow issues, and found one: it
> assumes that inotify cookies fit into an Emacs fixnum.

Thanks.  This needs to be fixed, indeed.

> This assumption isn't true on 32-bit hosts, unless Emacs is configured
> with --with-wide-int.  As you know I'm a fan of wide integers, and my
> preferred solution would be to make --with-wide-int the default, which
> would solve the problem.

No, that would only solve the problem if the --with-wide-int option is
removed on 32bit systems and imposed as the only choice.  Even if we may
change the default for 24.2 we're definitely not going to drop support
for "narrow int" operation so soon.


        Stefan