On Thu, Aug 11, 2016 at 01:04:09PM +0200, Francois Gouget wrote: > Servers that recognize this special report then stop streaming (sending > regular screen updates instead) while older ones essentially ignore it. > > Signed-off-by: Francois Gouget <fgouget@xxxxxxxxxxxxxxx> > --- > > This patchset is based on Victor Toso's idea [1] of using the stream > reports to tell the server that, despite expectations, the client cannot > handle a given stream. > > You'll notice that this patch does not directly check for > create_xxx_decoder() errors. Instead it relies on the previous patchset > [2] deleting broken streams so that the following messages will run into > an unknown stream. I can't help but wonder why it's better this way? Is it so that you can reuse the stream report code as mentioned below? > > Of course this could already happen in case of a malicious server > sending garbage to the client. So this patchset is quite independent > from the previous one. > > I don't know what the consequences of receiving an unknown message would > be for the server so I chose to hook into the > display_handle_stream_activate_report() as that's where we get notified > that the server supports the stream reports. Have you considered using a dedicated client -> server message for that rather than using magic values in a stream report message? Christophe
Attachment:
signature.asc
Description: PGP signature
_______________________________________________ Spice-devel mailing list Spice-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/spice-devel