thought I start a new thread the old one did not really have much meat to
it. So I got my dual head agp vga card. Couple things off the bat I think are odd: 1. VW only knows of one screen , so any math which assumes the physical monitor a particular window resides in will not work i.e. the monitor being the screen. Instead there's one virtual screen spread across the two monitors. So code that does stuff like "Screen default bounds center" assuming that the Screen is the monitor one is on will be unhappy. Is it reasonable to know what monitor one is one? I got around this but not 2. Very, very, very ugly is the fact that the splash screen straddles the two monitors :( 3. Have not had a chance to mimic Citrix via setting up terminal services on our Win 2000 server box. More later ... -- Charles A. Monteiro http://wiki.nycsmalltalk.org http://www.monteirosfusion.com http://monteirofusion.blogspot.com |
Screen object actually has a concept of subregions for each of your monitors and things like determining window positions all respect the setup by going through #boundsAround: and you should do too if you want to get down and dirty. I've had some good success in the past with dual monitor setups without any issues and herald is disabled in 7.5 by default anyway although I could swear it used to work normally too. Hope this helps, |
Boris:
thanks, I did not know about the subRegions. In 7.3 it does not pick up that I have multiple monitors i.e. multiple regions whereas in 7.4.1 it does. I think that the API surrounding subRegions could be made "nicer" but I can do as much here. I wonder if it makes sense to reify the notion of the "monitor" i.e. that specific piece of screen real estate viewed thru a physical monitor. We don't want to diasble the splash screen, we just don't think it makes sense that it straddles two monitors. I guess in the case of three monitors it will at least not straddle monitors, in the case of 4 monitors which again our clients have it may :) . Really, the vm should pop the splash screen offset from center of the primary monitor. As far as 7.5. I guess you are talking dev builds, unless I just have not received our copy yet. We won't provide our customers even if just to QA builds based on "unofficial" Cincom Smalltalk releases. So I have been waiting for the "imminent" release of 7.5 for quite a while. Not to start another thread here but for one I hope 7.5 takes care of the "stuck cursor" issue. We really need to clean up the "embarrassing" issues that get in the way of new sales. Issues like stuck cursors, splash screens that get split in half, drop downs popping up in the wrong screens in the context of Citrix which is pretty much a corporate standard for what it does, need to be cleaned up asap. -Charles On Tue, 10 Apr 2007 01:36:15 -0400, Boris Popov <[hidden email]> wrote: > Screen object actually has a concept of subregions for each of your > monitors and things like determining window positions all respect the > setup by going through #boundsAround: and you should do too if you want > to get down and dirty. I've had some good success in the past with dual > monitor setups without any issues and herald is disabled in 7.5 by > default anyway although I could swear it used to work normally too. Hope > this helps, > > Cheers! > > -Boris > (Sent from a BlackBerry) > > ----- Original Message ----- > From: Charles A. Monteiro <[hidden email]> > To: [hidden email] <[hidden email]> > Cc: Sean Glazier <[hidden email]> > Sent: Mon Apr 09 12:06:49 2007 > Subject: more dual monitor woes > > thought I start a new thread the old one did not really have much meat to > it. So I got my dual head agp vga card. Couple things off the bat I think > are odd: > > 1. VW only knows of one screen , so any math which assumes the physical > monitor a particular window resides in will not work i.e. the monitor > being the screen. Instead there's one virtual screen spread across the > two > monitors. So code that does stuff like "Screen default bounds center" > assuming that the Screen is the monitor one is on will be unhappy. > > Is it reasonable to know what monitor one is one? > > I got around this but not > > 2. Very, very, very ugly is the fact that the splash screen straddles the > two monitors :( > > 3. Have not had a chance to mimic Citrix via setting up terminal services > on our Win 2000 server box. > > More later ... > > > > > -- Using Opera's revolutionary e-mail client: http://www.opera.com/mail/ |
Hi,
I just found this very old post and wondered if the issue concerning drop down menus popping up on the wrong screen was fixed already? This happens to quite a few of our customers and is very embarrassing. A setup of 2 to 4 monitors is very common in video and music production studios. I vaguely remember something related to this was posted a while ago, but searching the list by keywords did not lead to any result. If by any chance someone can remember better, any pointer is highly appreciated. Thanks! Andre On 10.04.2007, at 16:29, Charles A. Monteiro wrote: > Re: more dual monitor woes > We really need to clean up the "embarrassing" issues that get in the > way of new sales. Issues like stuck cursors, splash screens that get > split in half, drop downs popping up in the wrong screens in the > context of Citrix which is pretty much a corporate standard for what > it does, need to be cleaned up asap. _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
In reply to this post by OCIT
Hi,
I just found this very old post and wondered if the issue concerning drop down menus popping up on the wrong screen was fixed already? This happens to quite a few of our customers and is very embarrassing. A setup of 2 to 4 monitors is very common in video and music production studios. I vaguely remember something related to this was posted a while ago, but searching the list by keywords did not lead to any result. If by any chance someone can remember better, any pointer is highly appreciated. Thanks! Andre On 10.04.2007, at 16:29, Charles A. Monteiro wrote: > Re: more dual monitor woes > We really need to clean up the "embarrassing" issues that get in the > way of new sales. Issues like stuck cursors, splash screens that get > split in half, drop downs popping up in the wrong screens in the > context of Citrix which is pretty much a corporate standard for what > it does, need to be cleaned up asap. _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
(Sorry for the duplicate post)
Doing some experiments, I found out that the issue occurs when the primary monitor (left) is plugged into the secondary (right) outlet of the video card. i.e. Windows thinks the monitors are arranged 2, 1. Some indexing on a lower OS level seems to matter here. Interestingly enough, the issue can be fixed by simply reversing (or: reordering) the instance variable #subRegions of Screen. The effect is immediate. I tried to always sort the regions from left to right by their dimensions, to no avail. A solution is to tell the user to swap cables and rearrange the monitors in the Windows display properties. Whether that also works with 4 monitors, I don't know. Any idea? Andre _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
Andre Schnoor wrote:
> > Doing some experiments, I found out that the issue occurs when the > primary monitor (left) is plugged into the secondary (right) outlet of > the video card. i.e. Windows thinks the monitors are arranged 2, 1. > Some indexing on a lower OS level seems to matter here. > > Interestingly enough, the issue can be fixed by simply reversing (or: > reordering) the instance variable #subRegions of Screen. The effect is > immediate. > > I tried to always sort the regions from left to right by their > dimensions, to no avail. > > A solution is to tell the user to swap cables and rearrange the > monitors in the Windows display properties. Whether that also works > with 4 monitors, I don't know. > > Any idea? > > Andre > I just tried to reproduce this on my machine - without success. Could you give me some hints what I might do wrong, e.g. where does the menu appear, where should it appear, where is the parent window of the menu - is it on both screen or only one - which one? About seven years ago we had a case where a broken video card driver caused windows to appear at a wrong position when using multiple monitors. One fix was to update the driver, the other was to move the window after opening to the position where it was just opened. See the attached method. Ralf -- Ralf Propach, [hidden email] Tel: +49 231 975 99 38 Fax: +49 231 975 99 20 Georg Heeg eK (Dortmund) Handelsregister: Amtsgericht Dortmund A 12812 <?xml version="1.0"?> <st-source> <time-stamp>From VisualWorks®, 7.6 of March 3, 2008 on March 5, 2009 at 1:09:12 am</time-stamp> <methods> <class-id>Graphics.Window</class-id> <category>initialize-release</category> <body package="OS-Window System" selector="map">map "Make the receiver be the topmost window when expanded. If the receiver is not visible (mapped), map it. Make sure that the background color is current. Note that the map state of a window is orthogonal to the iconification state. If an unmapped, collapsed window is mapped, it will be mapped as an icon (if icons are supported by the platform window manager)." | proposedDisplayBox | proposedDisplayBox := self displayBox. self setHostBackgroundColor. self primMap. mapped := true. proposedDisplayBox = self displayBox ifFalse: [self displayBox: proposedDisplayBox]. self becomeActive</body> </methods> </st-source> _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
On 05.03.2009, at 10:15, Ralf Propach wrote:
> Andre Schnoor wrote: >> Doing some experiments, I found out that the issue occurs when the >> primary monitor (left) is plugged into the secondary (right) outlet >> of >> the video card. i.e. Windows thinks the monitors are arranged 2, 1. >> Some indexing on a lower OS level seems to matter here. >> >> Interestingly enough, the issue can be fixed by simply reversing (or: >> reordering) the instance variable #subRegions of Screen. The effect >> is >> immediate. >> >> I tried to always sort the regions from left to right by their >> dimensions, to no avail. >> >> A solution is to tell the user to swap cables and rearrange the >> monitors in the Windows display properties. Whether that also works >> with 4 monitors, I don't know. >> >> Any idea? >> >> Andre >> > Hi Andre, > > I just tried to reproduce this on my machine - without success. > Could you give me some hints what I might do wrong, e.g. > where does the menu appear, where should it appear, where is the > parent window of the menu - is it on both screen or only one - which > one? The pulldown menu appeared on the right monitor at the left border, while the parent window was on the left screen. That "left" monitor was plugged into the "right" plug of the video card (but arranged in display properties to be the left monitor). IIRC, it was the #boundsAround: message of Screen that did not deliver the correct rectangle in this situation. > About seven years ago we had a case where a broken video card driver > caused windows to appear at a wrong position when using multiple > monitors. > One fix was to update the driver, the other was to move the window > after opening to the position where it was just opened. I don't think this is related to the driver, because a simple reordering of the subRegions of Screen can solve it. Andre _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
Andre Schnoor wrote:
> On 05.03.2009, at 10:15, Ralf Propach wrote: >> Hi Andre, >> >> I just tried to reproduce this on my machine - without success. >> Could you give me some hints what I might do wrong, e.g. >> where does the menu appear, where should it appear, where is the >> parent window of the menu - is it on both screen or only one - which >> one? > > The pulldown menu appeared on the right monitor at the left border, > while the parent window was on the left screen. That "left" monitor was > plugged into the "right" plug of the video card (but arranged in > display properties to be the left monitor). > > IIRC, it was the #boundsAround: message of Screen that did not deliver > the correct rectangle in this situation. > > > I don't think this is related to the driver, because a simple > reordering of the subRegions of Screen can solve it. > the first and second monitor? What is the result of Win32SystemSupport new spiGetWorkArea ? Ralf -- Ralf Propach, [hidden email] Tel: +49 231 975 99 38 Fax: +49 231 975 99 20 Georg Heeg eK (Dortmund) Handelsregister: Amtsgericht Dortmund A 12812 _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
On 05.03.2009, at 15:18, Ralf Propach wrote:
> When the menu opens on the wrong monitor, what are the co-ordinates of > the first and second monitor? Second monitor: -1900 ... 0 First monitor: 0 ... 1900 (incl. Windows taskbar) This is my normal setup, i.e. I usually have a second monitor left of the main monitor. > What is the result of Win32SystemSupport new spiGetWorkArea? Unfortunately I can't verify this at this time. The hardware was meanwhile rearranged for other things. BTW: This also happens on the Mac (same setup as above). Submenus open on the right monitor, although their parent menu is on the left. Drop down menus are fine, but their submenus not. This renders them completely unusable. Andre _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
> BTW: This also happens on the Mac (same setup as above). Submenus open
> on the right monitor, although their parent menu is on the left. Drop > down menus are fine, but their submenus not. This renders them > completely unusable. Ha! Found that bug: In MacOSXMenuView>>openFromParentMenuItemView: replace "0" with "monitorRectangle left" in this snippet of code: "A Menu should be positioned inside the monitors bounds, and the beginning of the labels should be readable" menuWindow moveTo: (menuWindow globalOrigin x + intoMonitor x max: monitorRectangle left) @ (menuWindow globalOrigin y + intoMonitor y). That does the trick. I could imaging that there are other places where the code silently assumes that "0" is always the leftmost monitor coordinate. Andre _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
In reply to this post by Andre Schnoor
Andre Schnoor wrote:
> On 05.03.2009, at 15:18, Ralf Propach wrote: > >> When the menu opens on the wrong monitor, what are the co-ordinates of >> the first and second monitor? > > Second monitor: -1900 ... 0 > First monitor: 0 ... 1900 (incl. Windows taskbar) > > This is my normal setup, i.e. I usually have a second monitor left of > the main monitor. > >> What is the result of Win32SystemSupport new spiGetWorkArea? > > Unfortunately I can't verify this at this time. The hardware was > meanwhile rearranged for other things. > > BTW: This also happens on the Mac (same setup as above). Submenus open > on the right monitor, although their parent menu is on the left. Drop > down menus are fine, but their submenus not. This renders them > completely unusable. > > Andre > > I found a case where someone suggested that Screen>>insetForTaskbar: was to blame, because at one place subRegions first is hard-coded in there. But that cannot explain bugs on the Mac. Since you said before, that reordering the subRegions helps, I would suggest trying to reorder the tests in Screen>>boundsAround: , because the only idea I have right now is, that more than one subRegion responds with true to containsPoint:. However, I am not sure, which is the correct order for the tests. Ralf -- Ralf Propach, [hidden email] Tel: +49 231 975 99 38 Fax: +49 231 975 99 20 Georg Heeg eK (Dortmund) Handelsregister: Amtsgericht Dortmund A 12812 _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
In reply to this post by Andre Schnoor
Andre Schnoor wrote:
>> BTW: This also happens on the Mac (same setup as above). Submenus open >> on the right monitor, although their parent menu is on the left. Drop >> down menus are fine, but their submenus not. This renders them >> completely unusable. > > > Ha! Found that bug: > > In MacOSXMenuView>>openFromParentMenuItemView: replace "0" with > "monitorRectangle left" in this snippet of code: > > "A Menu should be positioned inside the monitors bounds, and the > beginning of the labels should be readable" > menuWindow moveTo: > (menuWindow globalOrigin x + intoMonitor x max: monitorRectangle left) > @ (menuWindow globalOrigin y + intoMonitor y). > > That does the trick. > > I could imaging that there are other places where the code silently > assumes that "0" is always the leftmost monitor coordinate. > > Andre > Ok, I have never tried that with the Mac OS X look, but now I can reproduce that - created AR 56533 Sub menus pop up on wrong screen with Aqua look But reordering subRegions does not help for this, so there is another bug somewhere. Ralf -- Ralf Propach, [hidden email] Tel: +49 231 975 99 38 Fax: +49 231 975 99 20 Georg Heeg eK (Dortmund) Handelsregister: Amtsgericht Dortmund A 12812 _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
In reply to this post by Andre Schnoor
On Mar 5, 2009, at 8:14 AM, Andre Schnoor wrote: > Ha! Found that bug: <useful stuff snipped> Thanks Andre. AR 56527 has been submitted, and has now been fixed into this week's build. Thanks for the helpful report, diagnosis, and fix. -- Travis Griggs Objologist My Other Machine runs OSX. But then... so does this one. _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
Travis Griggs wrote:
> On Mar 5, 2009, at 8:14 AM, Andre Schnoor wrote: > > >> Ha! Found that bug: >> > <useful stuff snipped> > > Thanks Andre. AR 56527 has been submitted, and has now been fixed into > this week's build. Thanks for the helpful report, diagnosis, and fix. We see a similar issue on a single screen where the menu appears in the top left corner when it should be elsewhere. Its very intermittent and we haven't noticed any pattern to it. Will your fix fix that too? cheers Steve _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
Free forum by Nabble | Edit this page |