This works around Qt4 QProcess getting stuck waiting for its
SIGCHLD signal handler to be called.

We don't really want to inherit the blocked signal mask from our parent.

This reverts commit 96c7f9c2.
This did not work with KDE 4.6 anyhow.

Kludge beurk beurk

Fix a busy loop and socket leak

Fix a busy loop and socket leak. Thanks to Santiago Gimeno for
pointing it out (Debian bug #607869)

This reverts commit a1b3978d.

It doesn't fix anything and it breaks new connections
Signed-off-by: Pierre Ynard <linkfanel@yahoo.fr>

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Oked-by: Laurent Aimar
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

pub   1024D/45C11CA7 2011-01-30 [expires: 2012-01-30]
      Key fingerprint = 160A 96C4 CA7F 8ABA D43C  C087 62B9 7F54 45C1 1CA7
uid                  VideoLAN Release Signing Key (2011)

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

This affects way too many code paths. Let KDE fix their own crap.

Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>

This fixes cases where the stream length is unknown

Otherwise, when sending requests while already receiving data
(e.g. seeking) responses are not waited for properly