1 |
On Mon, Oct 13, 2008 at 09:10:34AM +0200, Alan McKinnon wrote: |
2 |
> On Monday 13 October 2008 04:02:19 Iain Buchanan wrote: |
3 |
> |
4 |
> > [snip] |
5 |
> > |
6 |
> > > I've configured it with TwinView |
7 |
> > |
8 |
> > as in: |
9 |
> > Option "TwinView" "True" |
10 |
> |
11 |
> Yes. Some output : |
12 |
> |
13 |
> $ sudo grep -i -e xinerama -e twinview /var/log/Xorg.0.log |
14 |
> (**) Option "Xinerama" "1" |
15 |
> (**) Xinerama: enabled |
16 |
> (**) NVIDIA(0): Option "TwinView" "1" |
17 |
> (**) NVIDIA(0): Option "TwinViewXineramaInfoOrder" "DFP-0" |
18 |
> (**) NVIDIA(0): TwinView enabled |
19 |
> (II) Initializing built-in extension XINERAMA |
20 |
> |
21 |
> $ sudo grep -i -e xinerama -e twinview /etc/X11/xorg.conf |
22 |
> Option "Xinerama" "1" |
23 |
> Option "TwinView" "1" |
24 |
> Option "TwinViewXineramaInfoOrder" "DFP-0" |
25 |
> |
26 |
> |
27 |
> > > The viewports are aligned along the top edge |
28 |
> > |
29 |
> > you mean move the mouse up and it appears on the next screen? Don't you |
30 |
> > want them aligned left / right of each other? |
31 |
> |
32 |
> My description wasn't clear. I mean the screens are physically and logically |
33 |
> laid out like so: |
34 |
> |
35 |
> +------------------------------+ |
36 |
> | | | |
37 |
> | 1 | 2 | |
38 |
> | |-----------+ |
39 |
> +------------------+ |
40 |
> |
41 |
> 1 is the notebook screen |
42 |
> 2 is the external lcd |
43 |
> below 2 is dead space. The mouse works correctly. |
44 |
> |
45 |
> > > and the |
46 |
> > > panel/kicker/plasma/whatever on every desktop environment insists on |
47 |
> > > trying to stretch across both monitors, into dead space on the right hand |
48 |
> > > one. |
49 |
> > |
50 |
> > Sounds like you haven't compiled stuff with the xinerama USE flag. I |
51 |
> > put it in make.conf, and then did a emerge --newuse. |
52 |
> |
53 |
> OK, I did that. The packages that got rebuilt are: |
54 |
> |
55 |
> $ equery hasuse xinerama |
56 |
> [ Searching for USE flag xinerama in all categories among: ] |
57 |
> * installed packages |
58 |
> [I--] [ ~] x11-apps/xdpyinfo-1.0.3 (0) |
59 |
> [I--] [ ~] x11-libs/qt-3.3.8b (3) |
60 |
> [I--] [ ~] x11-libs/gtk+-2.14.3-r2 (2) |
61 |
> [I--] [ ~] x11-libs/qt-gui-4.4.2 (4) |
62 |
> [I--] [ ] x11-misc/engage-9999 (0) |
63 |
> [I--] [ ~] kde-base/ksplash-4.1.2 (4.1) |
64 |
> [I--] [ ~] kde-base/plasma-workspace-4.1.2 (4.1) |
65 |
> [I--] [ ~] kde-base/ksplashml-3.5.10 (3.5) |
66 |
> [I--] [ ~] kde-base/systemsettings-4.1.2 (4.1) |
67 |
> [I--] [ ~] kde-base/kwin-4.1.2 (4.1) |
68 |
> [I--] [ ~] kde-base/libplasma-4.1.2 (4.1) |
69 |
> [I--] [ ~] kde-misc/knetworkmanager-0.2.2_p20080528 (0) |
70 |
> [I--] [ ~] kde-misc/filelight-1.0-r1 (0) |
71 |
> [I--] [ ] media-libs/libsdl-1.2.13 (0) |
72 |
> [I--] [ ] media-libs/xine-lib-1.1.15-r1 (1) |
73 |
> [I--] [ ] net-libs/xulrunner-1.8.1.17 (1.8) |
74 |
> [I--] [ ~] media-sound/kid3-1.0 (0) |
75 |
> [I--] [ ~] media-sound/amarok-1.4.10-r1 (0) |
76 |
> [I--] [ ~] media-video/mplayer-1.0_rc2_p27725-r1 (0) |
77 |
> [I--] [ ] media-video/xine-ui-0.99.5-r1 (0) |
78 |
> [I--] [ ~] media-video/gxine-0.5.903 (0) |
79 |
> [I--] [ ~] app-cdr/k3b-1.0.5-r3 (0) |
80 |
|
81 |
tabletka ~ # equery hasuse xinerama | wc -l |
82 |
285 |
83 |
|
84 |
most of them are apps from kde-base/* (3.5.9), seems that it changed between |
85 |
3.5.9 and 3.5.10, plus iwndow managers like fluxbox, openbox... |
86 |
|
87 |
> |
88 |
> Seems like the only things that would affect kde-3 apps is qt-3.3.8b. |
89 |
> Plus x11-libs/libXinerama and x11-proto/xineramaproto (both latest unstable) |
90 |
> are installed. |
91 |
> |
92 |
> [snip] |
93 |
> |
94 |
> > > I'd appreciate some pros and cons feedback from the list before I embark |
95 |
> > > on a huge emerge -e world to include Xinerama support. |
96 |
> > |
97 |
> > Why would you do -e world? How about `emerge -uN world` The N being |
98 |
> > --newuse. or `emerge -vauDN world`. |
99 |
> |
100 |
> I was running |
101 |
> /bin/think --exaggerate --frustrated --logic-level -3 |
102 |
> when I typed that :-) |
103 |
> |
104 |
> > check out my blog for how I did it: |
105 |
> > |
106 |
> > http://nthrbldyblg.blogspot.com/2008/08/nvidia-xinerama-on-dell-m6300.html |
107 |
> |
108 |
> Nice blog :-) |
109 |
> |
110 |
> I'll fiddle some more with these tips later in the day, but first a conceptual |
111 |
> question: I read that huge collection of docs from nvidia-drivers, and |
112 |
> concluded that Xinerama and TwinView are fundamentally different and |
113 |
> incompatible. i.e. Xinerama starts with two classic X screens and joins them |
114 |
> in software to make one big display - an abstraction layer if you will. |
115 |
> TwinView rips out the guts of X, dispenses with the notion of separate |
116 |
> screens for a TwinView display and gives you one giant screen with no API for |
117 |
> an app to see how this big screen is composed. So, you either use Xinerama or |
118 |
> TwinView, but not both. |
119 |
> |
120 |
> Obviously, this understanding of mine is flawed. Which bit did I get wrong? |
121 |
|
122 |
Xinerama consists basically of two parts, the protocol to communicate the |
123 |
position/sizes of screen between the Xserver and the applications (which |
124 |
you usually get by enabling the xinerama use flag) and an xserver part |
125 |
(module?) that you can use to set up the screens. What you said is |
126 |
correct for the Xserver setup part... |
127 |
You use either xinerama setup to put together completely different |
128 |
displays (might be different cards, such as one nvidia, one ati, ...) |
129 |
or twinview in case of a dualhead nvidia setup. But both this setups use |
130 |
the xinerama protocol to let the apps/wm know the placement of the monitors. |
131 |
> |
132 |
> -- alan dot mckinnon at gmail dot com |
133 |
> |
134 |
> |
135 |
|
136 |
-- |
137 |
_ |
138 |
| |
139 |
YoYo () Siska |
140 |
=================== |
141 |
http://www.ksp.sk/ |