[01:38] <RAOF> robert_ancell: mir_surface_type_freestyle is your winner.
[01:38] <robert_ancell> RAOF, thanks!
[01:40] <RAOF> You'll just need to not draw shadows.
[01:40] <RAOF> Because, as a client, you can't do that correctly, so shouldn't try :P
[01:40] <robert_ancell> Are the current shadows client side?
[01:41] <RAOF> For GTK? Pretty sure they are.
[01:41] <RAOF> Or, at least, GTK-on-Shell does client-side shadows.
[01:41] <robert_ancell> I'm looking at what needs to be done to fix https://plus.google.com/+PopescuSorin/posts/Q2ExSx7K94c
[01:41] <robert_ancell> I'm looking at what needs to be done to fix https://plus.google.com/+PopescuSorin/posts/Q2ExSx7K94c
[01:41] <robert_ancell> I'm looking at what needs to be done to fix https://plus.google.com/+PopescuSorin/posts/Q2ExSx7K94c
[01:42] <robert_ancell> I'm looking at what needs to be done to fix https://plus.google.com/+PopescuSorin/posts/Q2ExSx7K94c
[01:42] <RAOF> There are a couple of things that need to be fixed there.
[01:42] <robert_ancell> blah
[01:42] <RAOF> mir_surface_type_freestyle will get you no decorations...
[01:42] <robert_ancell> ok, my client was showing non of that going through :)
[01:42] <RAOF> ...but then you'll have no way of moving the window :)
[01:42] <RAOF> :P
[01:42] <robert_ancell> RAOF, yeah, windows can't move themselves?
[01:42] <RAOF> Correct.
[01:43] <RAOF> There's planned API for it.
[01:43] <robert_ancell> OK
[01:43] <RAOF> (Plus, I don't know if U8 actually does anything different with type_freestyle yet)
[01:43] <robert_ancell> Yeah, I suspect not
[01:43] <robert_ancell> We should open up some bugs if not already there
[01:43] <RAOF> (API is basically - ‘hey, Mir! This mouse button press started a window drag. Kindly move me with it’)
[01:44] <RAOF> Yeah, certainly.
[01:45] <robert_ancell> RAOF, you don't know of a bug for the proposed surface moving API do you?
[01:45] <robert_ancell> I'm not seeing one
[01:46] <RAOF> No, sorry.
[01:46] <robert_ancell> OK, I'll open one
[01:47] <RAOF> Huh, and there's not even a trello card.
[01:47] <RAOF> This is one of those things that we all just know needs to happen, I guess ;)
[01:47] <robert_ancell> Yeah, there's a lot of those that are invisible unless you know about some random conversation that occurred
[01:47] <robert_ancell> Let me shine some light :)
[01:50] <robert_ancell> RAOF, Does X have a clever way of telling the X server to "move with motion events" or does the client just convert every motion event into a window move request?
[01:50] <robert_ancell> i.e. are we improving on the status quo here?
[01:50] <RAOF> We're improving on the status quo here.
[01:50] <robert_ancell> Do you know if Wayland it doing the same?
[01:51] <RAOF> Yes; as long as by “Wayland” you mean GNOME Shell :)
[01:52] <RAOF> (Or, less snarkily, xdg-shell, but that's only implemented by weston and Shell to my current knowledge)
[01:52] <robert_ancell> Yes, I mean xdg-shell
[01:52] <RAOF> See the “move” and “resize” requests on xdg_surface
[01:52] <robert_ancell> Ah, bug 1398849 is what I was looking for
[01:56] <duflu> robert_ancell: The dragging functionality you want is logged as: https://bugs.launchpad.net/mir/+bug/1420334
[01:57] <duflu> Good news! Not only is 0.14.0 released in wily but we now get free dbgsym packages and more platforms than we test even :)
[01:59] <duflu> RAOF: Interestingly we build ppc archs that supposedly are big endian (by default). I bet they don't work...
[02:00] <RAOF> duflu: You are correct.
[02:00] <RAOF> There's only one spot that they don't work, though, which is pretty neat.
[02:00] <duflu> RAOF: The spot I documented?
[02:01] <RAOF> I don't know which spot you documented, but it was the spot in GLRenderer (IIRC?) that checks endianness and throws on BE architectures.
[02:02] <RAOF> Because we're too lazy to work out how to read big endian buffers.
[02:02] <duflu> RAOF: I documented what was missing for big endian in our pixel format typedef and also in ShmBuffer
[02:02] <duflu> That might be what you mean by GLRenderer
[02:03] <RAOF> Ah, GlPixelBuffer.
[02:44] <RAOF> Ok. Thanks, internet outage, for getting me down to Hillarys to test out the Ubuntu Touch wifi hotspot silo...
[13:55] <guest42345> hm firefox nightly builds are on gtk3
[14:58] <attente> for a menu surface with mir_edge_attachment_horizontal, how does mir know to align the left edges (for LTR text) or the right edges (for RTL edges)?
[14:58] <attente> *RTL text
[14:59] <alan_g> attente: there's logic in CanonicalWindowManagerPolicy for that placement. (If that's what you're asking)
[15:01] <attente> alan_g: how does a mir client change this behaviour?
[15:02] <alan_g> A mir client doesn't - it request a placement and the server chooses how to implement it
[15:05] <attente> alan_g: ok, thanks