Re: Pretest next week

[YAMAMOTO Mitsuharu]

>>>>> On Wed, 4 Feb 2009 12:08:55 +0000 (UTC), Adrian Robert <address@hidden> 
>>>>> said:

>> I thought it's shameful for Emacs to include a port that doesn't
>> handle C-g properly as a part of its official release.  Also, I was
>> not suggesting the removal of the port from the whole CVS branches.

> I believe it is working 100% now.  Please report any specific
> failure cases you find through report-emacs-bug

> The cost of the implementation has been that are some cases where
> menus are not updated when clicked on; specifically, when
> ns_read_socket is called asynchronously through handle_async_input
> or poll_for_input_1 (depending if SYNC_INPUT is enabled), the menu
> update must be deferred.

> I believe the effect of this on user experience will be rare to
> almost nonexistent in practice, but let's find out empirically.  If
> it is a problem there are a couple of approaches that could be
> tried, but I'd rather prioritize other issues for now unless it
> proves to be a problem in actual use.

What happens if read_socket_hook is called from the QUIT macro in the
context of process filters or idle timers?

                                     YAMAMOTO Mitsuharu
                                address@hidden