starting mythfrontend is very verbose

[JYA]

On a mac, running master.

Start mythfrontend

Log will fill with :

2013-04-30 13:35:08.088384 E  Socket readBlock error 0
2013-04-30 13:35:08.088720 E  Socket readBlock error 9
2013-04-30 13:35:08.088756 E  Socket readBlock error 9
2013-04-30 13:35:08.088912 E  Socket readBlock error 9
2013-04-30 13:35:08.088967 E  Socket readBlock error 9
2013-04-30 13:35:08.089168 E  Socket readBlock error 9
2013-04-30 13:35:08.089236 E  Socket readBlock error 9
2013-04-30 13:35:08.089418 E  Socket readBlock error 9
2013-04-30 13:35:08.089581 E  Socket readBlock error 9
2013-04-30 13:35:08.089752 E  Socket readBlock error 9
2013-04-30 13:35:08.089890 E  Socket readBlock error 9
2013-04-30 13:35:08.090034 E  Socket readBlock error 9
2013-04-30 13:35:08.092493 E  Socket readBlock error 9
2013-04-30 13:35:08.114291 E  Socket readBlock error 9

[danielk]

According to the Error enum here:

​http://www.tbi.univie.ac.at/~pmg/tutorials/QT/html/qsocketdevice.html

"Socket readBlock error 0" is NoError and "Socket readBlock error 0" is UnknownError.

The debugging + nearby code was introduced in [44df943b0f9]. This is independent of the MythSocket code and can be reverted independently if I can't find the bug.

It is very weird that ret would be < 0 and then error returns 0.

[JYA]

After tracking.. The issue comes from ::readrecv returning -1 and setting errnor = EAGAIN

so MSocketDevice::readData returns -1 as if an error occurred, when really it should only try to read again (seeing the it's called from readBlock()

why would the socket indicates that it has XX bytes available; but returns EAGAIN; I'm not sure..

But this needs addressing in MythSocket

[Jean-Yves Avenard <jyavenard@…>]

In 8882c7551845a302b53e5d7a0bfb0fcbd929c207/mythtv:

Error: Processor CommitTicketReference failed

GIT backend not available