Re: Trellis WXEventLoop patch

On Wed, Jun 18, 2008 at 04:15, Phillip J. Eby <pje-Wh6+Hckhi6HFNGf7iClzIwC/[email protected]> wrote:
> Looking at the patch, it appears you are trying to support calling 'run()'
> from *inside* the wx main loop, but I don't currently have any plans to
> support that feature: not with Twisted, and definitely not with wx.  I might
> consider adding support for it in the future, however.  (In the meantime, of
> course, you can simply use your own subclass of WXEventLoop to support
> this.)

Sure, that's what I'm doing right now.


> As for the other part of your patch, did you encounter some *actual* problem
> with CallAfter not accepting a zero delay?  For that matter, did you
> encounter a situation where self._next_time was actually zero?  I'm curious,
> because it seems to me that zero is the one value that self._next_time
> should NEVER be able to be.  If you can provide a test that reproduces that
> situation (with or without wx), I would appreciate it.

I knew CallAfter doesn't accept zero delay, so first I added checks
for that. But it didn't help because as it turned out I was seeing
negative _next_time values. If that sounds like a bug, I think I know
what minimal wx app will reproduce the problem.


-- 
Best Regards,
Sergey Schetinin

http://s3bk.com/ -- S3 Backup
http://word-to-html.com/ -- Word to HTML Converter