[mythtv] Ticket #4270: MythTV on Windows (with MinGW)

[buzz]

Andrei,
you mentioned a little whilke back that QSocketNotifier was never generating
signals, I was reviewing the doco (
http://doc.trolltech.com/3.3/qsocketnotifier.html
 ) and it says:
  " The behavior of a write notifier that is left in enabled state after
having emitting the first activated() signal (and no "would block" error has
occurred) is undefined. Depending on the operating system, it may fire on
every pass of the event loop or not at all. "

So based on this information, I's suggest that it's possible that there is
somewhere a "write notifier that is left in enabled state" which is
resulting in the notifier firing "not at all".  (which is one of the
platform-specific possibilities that are not defined)

what do u think, possible?

Am I stating the obvious?...   this QT stuff is rather foreign to me.....

Buzz.

On Dec 29, 2007 4:00 PM, Andrei Tanas <andrei at tanas.ca> wrote:

> I have one more critical QT patch - I was trying to get backend working
> (successfully, the next step will be to implement a DirectShow-based
> recorder), when I realized that I hit another QT bug. Since QMutex is used
> all over the place, fixing this bug seems to have made the frontend more
> stable too (!).
>
> Since it's not related to MythTV itself, I don't think I should post it to
>
> trac, but I think it's important to post it somewhere.
>
> I will update the wiki and post a URL for download it.
>
> Sorry guys, looks like you'll have to rebuild QT again (full recompile is
> not necessary, but linking takes hours, somebody needs to fix that MinGW
> linker...).
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mythtv.org/pipermail/mythtv-dev/attachments/20071229/186abf52/attachment.htm