Testing out F13-Beta-64. I was puzzled why I could only see my mouse cursor and no Anaconda on the monitor. Obviously the x-server was loaded.
Then I realized that even though my DVI connected TV was powered *off* it was selected as primary screen, instead of my CRT (which was powered on). So Anaconda was on another screen.
In some ways I am happy that dual-head works out of the box (nouveau), but I had to physically unplug my DVI cable before I could get my CRT to be used for installation.
My question is: Is this the correct behavior we want? If so, how could I tell Anaconda to use my CRT as primary?
I'm hesitant to call this a bug.
I plan on running many more installs before F13 is released. This will be incredibly annoying.
On Wed, Apr 14, 2010 at 2:12 AM, Mauriat Miranda mirandam@mjmwired.com wrote:
Testing out F13-Beta-64. I was puzzled why I could only see my mouse cursor and no Anaconda on the monitor. Obviously the x-server was loaded.
Then I realized that even though my DVI connected TV was powered *off* it was selected as primary screen, instead of my CRT (which was powered on). So Anaconda was on another screen.
In some ways I am happy that dual-head works out of the box (nouveau), but I had to physically unplug my DVI cable before I could get my CRT to be used for installation.
My question is: Is this the correct behavior we want? If so, how could I tell Anaconda to use my CRT as primary?
I'm hesitant to call this a bug.
This *is* a bug, so fill it.
----- "drago01" drago01@gmail.com wrote:
On Wed, Apr 14, 2010 at 2:12 AM, Mauriat Miranda mirandam@mjmwired.com wrote:
Testing out F13-Beta-64. I was puzzled why I could only see my
mouse
cursor and no Anaconda on the monitor. Obviously the x-server was loaded.
Then I realized that even though my DVI connected TV was powered
*off*
it was selected as primary screen, instead of my CRT (which was powered on). So Anaconda was on another screen.
In some ways I am happy that dual-head works out of the box
(nouveau),
but I had to physically unplug my DVI cable before I could get my
CRT
to be used for installation.
My question is: Is this the correct behavior we want? If so, how could I tell Anaconda to use my CRT as primary?
I'm hesitant to call this a bug.
This *is* a bug, so fill it.
It is very unfortunate that displays are not mirrored in installation process any more. There is no reason to have spanned desktop when installing system. I even asked anaconda developers some time ago. I don't remember the answer precisely, but I guess it was something like it was too complex to change it just for the installation process.
On Wed, Apr 14, 2010 at 8:34 AM, Kamil Paral kparal@redhat.com wrote:
It is very unfortunate that displays are not mirrored in installation process any more. There is no reason to have spanned desktop when installing system. I even asked anaconda developers some time ago. I don't remember the answer precisely, but I guess it was something like it was too complex to change it just for the installation process.
Are you referring to this? https://bugzilla.redhat.com/show_bug.cgi?id=521803
Is the anaconda Xorg log saved off, or do I need to capture it in a terminal during install?
----- "Mauriat Miranda" mirandam@mjmwired.com wrote:
On Wed, Apr 14, 2010 at 8:34 AM, Kamil Paral kparal@redhat.com wrote:
It is very unfortunate that displays are not mirrored in
installation
process any more. There is no reason to have spanned desktop when
installing system. I even asked anaconda developers some time ago.
I don't remember the answer precisely, but I guess it was something like it was too complex to change it just for the installation process.
Are you referring to this? https://bugzilla.redhat.com/show_bug.cgi?id=521803
No, that's just another problem caused by spanned desktop.
Is the anaconda Xorg log saved off, or do I need to capture it in a terminal during install?
AFAIK it's not saved to target system, but you can find it during the installation in /tmp I guess.
On Wed, 2010-04-14 at 08:34 -0400, Kamil Paral wrote:
It is very unfortunate that displays are not mirrored in installation process any more. There is no reason to have spanned desktop when installing system. I even asked anaconda developers some time ago. I don't remember the answer precisely, but I guess it was something like it was too complex to change it just for the installation process.
That's basically it, yeah. There's no very simple way to specify clone mode instead of spanned mode for X, X just doesn't provide a simple way to do it. We'd have to create an xorg.conf file only to be used during Anaconda, I think that'd be the simplest way to do it, and doing that isn't terribly straightforward.
On Wed, Apr 14, 2010 at 7:56 PM, Adam Williamson awilliam@redhat.com wrote:
On Wed, 2010-04-14 at 08:34 -0400, Kamil Paral wrote:
It is very unfortunate that displays are not mirrored in installation process any more. There is no reason to have spanned desktop when installing system. I even asked anaconda developers some time ago. I don't remember the answer precisely, but I guess it was something like it was too complex to change it just for the installation process.
That's basically it, yeah. There's no very simple way to specify clone mode instead of spanned mode for X, X just doesn't provide a simple way to do it.
There is an extension called XRandR ... so no it is possible without creating an xorg.conf (you can enable and disable it at runtime).
On Wed, 2010-04-14 at 20:01 +0200, drago01 wrote:
On Wed, Apr 14, 2010 at 7:56 PM, Adam Williamson awilliam@redhat.com wrote:
On Wed, 2010-04-14 at 08:34 -0400, Kamil Paral wrote:
It is very unfortunate that displays are not mirrored in installation process any more. There is no reason to have spanned desktop when installing system. I even asked anaconda developers some time ago. I don't remember the answer precisely, but I guess it was something like it was too complex to change it just for the installation process.
That's basically it, yeah. There's no very simple way to specify clone mode instead of spanned mode for X, X just doesn't provide a simple way to do it.
There is an extension called XRandR ... so no it is possible without creating an xorg.conf (you can enable and disable it at runtime).
that's roughly equally inelegant, though also a possibility.
On Wed, 2010-04-14 at 11:21 +0200, drago01 wrote:
On Wed, Apr 14, 2010 at 2:12 AM, Mauriat Miranda mirandam@mjmwired.com wrote:
Testing out F13-Beta-64. I was puzzled why I could only see my mouse cursor and no Anaconda on the monitor. Obviously the x-server was loaded.
Then I realized that even though my DVI connected TV was powered *off* it was selected as primary screen, instead of my CRT (which was powered on). So Anaconda was on another screen.
In some ways I am happy that dual-head works out of the box (nouveau), but I had to physically unplug my DVI cable before I could get my CRT to be used for installation.
My question is: Is this the correct behavior we want? If so, how could I tell Anaconda to use my CRT as primary?
I'm hesitant to call this a bug.
This *is* a bug, so fill it.
I disagree. The driver cannot tell whether a connected display is powered on or off, and there is no method for choosing which display is 'primary' which will always be correct (why should we choose a VGA-connected display over a DVI-connected one, or vice versa? There's no way of knowing which is the 'primary' display in any particular case).
On Wed, Apr 14, 2010 at 7:55 PM, Adam Williamson awilliam@redhat.com wrote:
On Wed, 2010-04-14 at 11:21 +0200, drago01 wrote:
On Wed, Apr 14, 2010 at 2:12 AM, Mauriat Miranda mirandam@mjmwired.com wrote:
Testing out F13-Beta-64. I was puzzled why I could only see my mouse cursor and no Anaconda on the monitor. Obviously the x-server was loaded.
Then I realized that even though my DVI connected TV was powered *off* it was selected as primary screen, instead of my CRT (which was powered on). So Anaconda was on another screen.
In some ways I am happy that dual-head works out of the box (nouveau), but I had to physically unplug my DVI cable before I could get my CRT to be used for installation.
My question is: Is this the correct behavior we want? If so, how could I tell Anaconda to use my CRT as primary?
I'm hesitant to call this a bug.
This *is* a bug, so fill it.
I disagree. The driver cannot tell whether a connected display is powered on or off
Depends on the display, most displays are only detectable when they are actually powered on.
, and there is no method for choosing which display is 'primary'
Sure there is .... xrandr provides a way to choose a primary display, so worst case ask the user, or in case of the install just clone the output.
On Wed, 2010-04-14 at 19:59 +0200, drago01 wrote:
I disagree. The driver cannot tell whether a connected display is powered on or off
Depends on the display, most displays are only detectable when they are actually powered on.
That's not the case at all in my experience. Clearly in this case it's not true, otherwise it wouldn't know the TV was there at all.
, and there is no method for choosing which display is 'primary'
Sure there is .... xrandr provides a way to choose a primary display, so worst case ask the user, or in case of the install just clone the output.
That's not what I meant. I meant that there's no reliable heuristic for knowing which is primary.
Ask the user? How do you pick which display to ask the user on? :)
On Wed, Apr 14, 2010 at 8:05 PM, Adam Williamson awilliam@redhat.com wrote:
On Wed, 2010-04-14 at 19:59 +0200, drago01 wrote:
I disagree. The driver cannot tell whether a connected display is powered on or off
Depends on the display, most displays are only detectable when they are actually powered on.
That's not the case at all in my experience. Clearly in this case it's not true, otherwise it wouldn't know the TV was there at all.
Well the number of broken connector tables is infinite ... my laptop does detect that a TV is connected even though there is no TV connector at all (one has to use a docking station to be able to connect one).
, and there is no method for choosing which display is 'primary'
Sure there is .... xrandr provides a way to choose a primary display, so worst case ask the user, or in case of the install just clone the output.
That's not what I meant. I meant that there's no reliable heuristic for knowing which is primary.
Ask the user? How do you pick which display to ask the user on? :)
Both ;) ... but anyway cloning would be the right solution here, anaconda is limited to 800x600 anyway so there is no need for a multiscreen view.
On Wed, 2010-04-14 at 20:11 +0200, drago01 wrote:
On Wed, Apr 14, 2010 at 8:05 PM, Adam Williamson awilliam@redhat.com wrote:
On Wed, 2010-04-14 at 19:59 +0200, drago01 wrote:
I disagree. The driver cannot tell whether a connected display is powered on or off
Depends on the display, most displays are only detectable when they are actually powered on.
That's not the case at all in my experience. Clearly in this case it's not true, otherwise it wouldn't know the TV was there at all.
Well the number of broken connector tables is infinite ... my laptop does detect that a TV is connected even though there is no TV connector at all (one has to use a docking station to be able to connect one).
True TV outputs are known to be problematic, yes, but the TV in this case was connected by DVI, which is far less likely to be broken. In my experience, drivers usually detect monitors that are connected by D-SUB/DVI but powered off.
Both ;) ... but anyway cloning would be the right solution here, anaconda is limited to 800x600 anyway so there is no need for a multiscreen view.
I agree, but as I said, I already proposed this back in the F12 cycle and the anaconda developers considered it too messy to implement. By all means try again, if you like.
On Wed, Apr 14, 2010 at 2:51 PM, Adam Williamson awilliam@redhat.com wrote:
On Wed, 2010-04-14 at 20:11 +0200, drago01 wrote:
Well the number of broken connector tables is infinite ... my laptop does detect that a TV is connected even though there is no TV connector at all (one has to use a docking station to be able to connect one).
True TV outputs are known to be problematic, yes, but the TV in this case was connected by DVI, which is far less likely to be broken. In my experience, drivers usually detect monitors that are connected by D-SUB/DVI but powered off.
Both ;) ... but anyway cloning would be the right solution here, anaconda is limited to 800x600 anyway so there is no need for a multiscreen view.
I agree, but as I said, I already proposed this back in the F12 cycle and the anaconda developers considered it too messy to implement. By all means try again, if you like.
So, NOTABUG or CANTFIX then I guess.
My workaround for now is using 'vesa', which is acceptable for Anaconda. After install, I will have to setup my Xorg.conf, which most people do for dual-head anyways.
On 14/04/10 04:54 PM, Mauriat Miranda wrote:
My workaround for now is using 'vesa', which is acceptable for Anaconda. After install, I will have to setup my Xorg.conf, which most people do for dual-head anyways.
No need for xorg.conf if you can use xrandr for dual head...
----- "Mauriat Miranda" mirandam@mjmwired.com wrote:
Testing out F13-Beta-64. I was puzzled why I could only see my mouse cursor and no Anaconda on the monitor. Obviously the x-server was loaded.
Then I realized that even though my DVI connected TV was powered *off* it was selected as primary screen, instead of my CRT (which was powered on). So Anaconda was on another screen.
In some ways I am happy that dual-head works out of the box (nouveau), but I had to physically unplug my DVI cable before I could get my CRT to be used for installation.
My question is: Is this the correct behavior we want? If so, how could I tell Anaconda to use my CRT as primary?
I'm hesitant to call this a bug.
I plan on running many more installs before F13 is released. This will be incredibly annoying.
I believe I have accidentally just reported very similar (or the same) bug: https://bugzilla.redhat.com/show_bug.cgi?id=582525
In case powered off displays are really detected as powered on displays (as some people claim here, I didn't know that) and it can't be fixed, then IMHO having spanned desktop by default is a very bad idea...