andi- has quit [Remote host closed the connection]
andi- has joined #spectrum
ddima_ has joined #spectrum
ddima has quit [Ping timeout: 245 seconds]
<edef>
qyliss: did we have a reason for using xpra over Xephyr?
<edef>
just rootlessness?
<qyliss>
What transport does Xephyr use?
<edef>
it renders to a framebuffer, exposed as an SHM XImage
<edef>
i think i could plausibly uncouple those two parts
<edef>
otoh multiple xwayland instances are a pretty viable strategy as well
<qyliss>
So would I have to give something running in isolation my full frame buffer?
<qyliss>
Yeah
<qyliss>
Xwayland is probably the way
<qyliss>
Or, well, ideally just wayland
<edef>
20:20:30 < qyliss> So would I have to give something running in isolation my full frame buffer?
<edef>
no, this is a misparse
<edef>
you are assuming there is only one frame buffer
<edef>
wayland relies on the same principle
<qyliss>
oh, right
<qyliss>
Gotcha
<edef>
every application renders to a frame buffer in shared memory
<edef>
shared with the compositor
MichaelRaskin has joined #spectrum
<MichaelRaskin>
Does wayland have enough network transparency for that?
<MichaelRaskin>
My own stupid story: I am too lazy to setup Xpra all-around, so I do the bad thing of giving programs access to my X session so far… normally proxying the socket into the jail is enough, but Flash Player really wants shared buffer (and so shared IPC space)
<qyliss>
MichaelRaskin: you can use Wayland over a network with a little extra software
<qyliss>
The very first thing I did for the project that could now be called Spectrum, >1 year ago, was a 100 line Nix function that wrapped a program in a container and did Xpra to the host