Code review comment for lp:~mir-team/mir/vsync-timings-from-server

First pass:

Could VsyncProvider be in mir::graphics instead of mir::frontend?

new whitespace
324 +

int64_t mir_surface_get_last_display_time(MirSurface *surface)
The naming has some interesting implications... like could the client api user rely on this to roughly see how long the screen has been off? Also, I'd think that the vsync stuff would be in mir_connection with the rest of the display stuff.

599 + virtual std::chrono::nanoseconds last_vsync_for(graphics::DisplayConfigurationOutputId output) = 0;
and
204 + virtual std::shared_ptr<frontend::VsyncProvider> the_vsync_provider();
and
184 + virtual std::shared_ptr<frontend::VsyncProvider> vsync_provider() = 0;

Seems like a lot of rigging where we could just have
mg::Display::last_vsync_for(graphics::DisplayConfigurationOutputId output) = 0;
and the server code could just use the existing the_display() function.