[mythtv] Autodiscovery Patch for MythFrontend.
devel at mrwire.co.uk
Sun Apr 22 14:31:47 UTC 2007
This is working great on two machines here. One is a combined FE/BE,
the other FE only.
One thing I'm unable to test is the security pin stuff since I've no
idea where I would setup the pin in the first place?
On Sun, 2007-04-08 at 11:33 -0400, David wrote:
> Attached is a patch against Revision 13184 that adds Auto-Discovery to the
> NOTE: The Screens to pick a backend and enter a password are downright UGLY!
> The code in the patch is completely functional, and only needs cosmetic
> A few things to note:
> - Themes don't work without a database available for settings. This
> includes the existing dialog that should be presented to the user if
> mysql.txt is missing and the defaults don't work. I was able to trick it
> into working by calling OverrideSettingForSession with 31 settings, however
> as soon as the MythMainWindow is created, it tries to insert all of the
> global key bindings into the database which fails. Although I could have
> continued down this route, it seemed like too much of a hack so I scrapped
> it for the current implementation.
> - I tried to make as few changes to mythcontext as possible. If the user
> chooses "No master Backend" it will default to the existing program flow.
> - A new Command Line option -p or --prompt will force the selection screen
> to be displayed even if the config.xml file has a valid USN in it.
> - The autodiscovery code stores the users default selection in the
> .mythtv/config.xml file. It's also the file that can be used to override
> default UPnP settings. (See prior commit for more details on file format).
> - Only the USN (Unique Service Name) of the backend is stored, not it's IP
> address. This allows the backend to change IPs at will, and the frontend
> will always be able to connect to it with no user intervention.
> - All of the poorly designed dialog boxes are in the masterselection.cpp/.h
> files in case anyone wants to improve upon them.
> - With this patch, the frontend starts to participate in the UPnP protocol
> as a AVMediaRenderer. A full implementation at this time would make the
> patch too complex. If this, or a modified version of this patch gets
> committed, I will add the remaining classes to stub out all services needed
> for a fully compliant UPnP Renderer. (However, this patch does make the
> frontend a discoverable device).
> Let me know if anyone has any questions.
> David Blain.
> mythtv-dev mailing list
> mythtv-dev at mythtv.org
More information about the mythtv-dev