We've got some in-house applications built in MFC, with OpenGL drawing routines. They all use the same code to draw on the screen and either print the screen or save it to a JPEG file. Everything's been working fine in Windows XP, and I need to find a way to make them work on Vista.
In three of our applications, everything works. In the remaining one, I can get the window border, title bar, menus, and task bar, but the interior never shows up. As I said, these applications use the exact same code to write to the screen and capture the window image, and the only difference I see that looks like it might be relevant is that the problem application uses the MFC multiple document interface, while the ones that work use the single document interface.
Either the answer isn't on the net, or I'm worse at Googling than I thought. I asked on the MSDN forums, and the only practical suggestion I got was to use GDI+ rather than GDI, and that did nothing different. I have tried different things with every part of the code that captures and prints or save, given a pointer to the window, so apparently it's a matter of the window itself. I haven't rebuilt the offending application using SDI yet, and I really don't have any other ideas.
Has anybody seen anything like this?
What I've got is four applications. They use a lot of common code, and share the actual .h and .cpp files, so I know the drawing and screen capture code is identical.
There is a WindowtoDIB() routine that takes a *pWnd, and a source rectangle and destination size. It looks like very slightly adapted Microsoft code, and I've found other functions in this file on the Microsoft website. Of my four applications, three handle this just fine, but one doesn't. The most obvious difference is that the problem one is MDI.
It looks to me like the *pWnd is the problem. I'm not a MFC guru by a long shot, and it seems to me that the problem may be that we've got one window setup in the SDIs, and more than one in the MDI. I may be passing the wrong *pWnd to the function.
In the meantime, it has started working properly on the 64-bit Vista test machine, although it still doesn't work on the 32-bit Vista machine. I have no idea why. I haven't changed anything since the last tests, and I didn't think anybody else had. (On the 32-bit version, the Print Screen key works as expected, but it does not save the screen as a JPEG.)
Your question title mentions screen capture but your actual question doesn't. Please elaborate more clearly. Is the problem that you can do screen capture of three of your applications, but not the fourth one? You can use different screen capture software that can capture OpenGL/DirectX windows. Those surfaces are handled directly by the Window Manager and won't show up with a simple 'PrtScn'.
Switching to GDI+ won't solve it, nor will switching to SDI.
If it's the content of the CView that you want, then yes, that should be right one. If it's the content of the whole screen (at least the content, without the toolbar(s) and status bar), then you should pass it the CMainFrame (that's the default name which may have been changed, the one that is derived from CMDIFrameWnd).
Can you post the code of WindowToDIB()? I've just tried it and It Works For Me (TM), but without OpenGL code in the view. Try passing the following windows to your WindowToDIB() function:
CMainFrame* mainfrm = static_cast<CMainFrame*>(::AfxGetMainWnd());
- mainfrm
- mainfrm->MDIGetActive()
- mainfrm->MDIGetActive()->GetActiveView()
and see what you get.
The contents of each window are directX surfaces and are only assembled by the window manager in the graphics card. You'd not be able to capture this unless you switch off the new interface (DWM) or code specifically for screen capture from the DWM.
Wikipedia has a good description of the Desktop Window Manager (DWM)
Sorry, I still don't understand. You're trying to get the Print Screen key to work on all four applications? Or you're trying to get the WindowtoDIB() function to work, which takes a 'screenshot' (from within your own application) of the application itself, so that it can be saved as an image file?
Also, what do you mean with 'he Print Screen key works as expected, but it does not save the screen as a JPEG.'? Print Screen only copies to the clipboard, what happens when you paste in Paint?
If your WindowtoDIB() function only 'captures' the window you pass to it, then yes, your MDI child windows are not going to show up.
We eventually solved this by creating a different OpenGL context, and drawing everything to that. We gave up on the screen capture.
Related
Some context
I've recently switched to ubuntu budgie (from unity), and I am really tired of the Plank/panel menu combo. I cannot find a setting that suits me, because depending on my screen setup, there's always something in the wrong place.
I am literally unable to show the menu on certain edges if I activate auto-hide, and if I don't activate it, it's not nice at all, to the point that I have removed the plank thing altogether. (Am I having strange bugs on this OS, or is it really messy?)
My idea
With great frustrations come new ideas. I thought again about one I had in the past. I would like to have a circle menu that pops around my mouse cursor when I press a given key combination (very much the kind of thing you would find in some games).
The main use case is to get "pined" application shortcuts easily when I need them, but perhaps other things would fit well with them (commands ...).
Questions
So my questions are:
Does such a thing already exist?
If it doesn't, is it difficult to realize? (How much time, complexity, ...)
What tools/libraries are needed for such a project? I know I'll find plenty of explanations on the gnome developer website but I could really use some more help.
Since you mention a buggy behaviour on Plank, depending on the screen configuration, I suspect you are suffering from this bug. In short: Plank's returned values for the space it needs are not always correct in multi monitor setup.
A neat option to replace at least part of the functionality is Ulauncher, by default called from a shortcut, but you could trigger it from anything that is capable of running its command.
Since Ulauncher's window simply identifies in the window list, you can easily write a script to move it to the current mouse position.
In case you'd need any help in that, just leave a comment.
Not sure if you are also referring to quick access of the window list, but for that you could use the Window Previews applet, or even the Workspace Overview applet, so life without Plank is possible.
I've got a bit of an interesting problem here. There are plenty of threads I've found where people are working to hide or get rid of a cursor on an embedded Qt GUI...but I'm trying to get a cursor to show up on an embedded Qt GUI.
I inherited a project that was 'finished' some time ago, and the person who did the most work on the project has moved on. Fast forward to today and there is a need to add a cursor to this functional touchscreen GUI. The system OS is Yocto Linux, and it is running a Qt 5.4 application on a framebuffer.
I've scoured the Qt code and there is nothing there that would hide a cursor. I've added in the appropriate QT_QPA_FB_HIDECURSOR=0 environment variable to my Qt startup script. I've experimented with adding a QCursor obejct to the GUI. Unfortunately none of these things are working. Using the QCusor I am sometimes able to get a cursor up on the screen, but isn't tied to the touch input (the cursor shows up at the position I programatically move it to, but it stays there when I interact with the GUI).
My touch input events are tied into Qt (via QT_QPA_GENERIC_PLUGINS=evdevtouch and QT_QPA_EVDEV_TOUCHSCREEN_PARAMETERS=/dev/input/event9:rotate=180), but for some reason that touch input cannot be tied to a cursor.
At this point I've spent a few days messing around with environment variables and startup script modifications, but nothing I've done has got the result I'm looking for.
Does anybody out there have some ideas on where to look for solutions to this problem?
Thanks!
Ian
So, now 3 months later I think my team and I just came up with a passable solution to this problem.
The path towards the solution started with the Qt Documentation on "Using libinput". The documentation boils down to a few important statements:
Parameters like the device node name can be set in the environment variables QT_QPA_EVDEV_MOUSE_PARAMETERS, QT_QPA_EVDEV_KEYBOARD_PARAMETERS and QT_QPA_EVDEV_TOUCHSCREEN_PARAMETERS
The mouse cursor shows up whenever QT_QPA_EGLFS_HIDECURSOR (for eglfs) or QT_QPA_FB_HIDECURSOR (for linuxfb) is not set and Qt's libudev-based device discovery reports that at least one mouse is available. When libudev support is not present, the mouse cursor always show up unless explicitly disabled via the environment variable.
The evdevtablet plugin provides basic support for Wacom and similar, pen-based tablets. It generates QTabletEvent events only. To enable it, pass QT_QPA_GENERIC_PLUGINS=evdevtablet in the environment or, alternatively, pass -plugin evdevtablet argument on the command-line. The plugin can take a device node parameter, for example QT_QPA_GENERIC_PLUGINS=evdevtablet:/dev/event1, in case the Qt's automatic device discovery (based either on libudev or a walkthrough of /dev/input/event*) is not functional or misbehaving.
So, in my system I have the device nodes: event0, event1, event2, event3, event4, event5, mice, and mouse0. Because I'm trying to get the mouse working, I made the assumption that I'd have to use the mouse0 node. This lead to me setting these environment variables:
QT_QPA_GENERIC_PLUGINS=evdevmouse
QT_QPA_EVDEV_MOUSE_PARAMETERS=/dev/input/mouse0
Much to my frustration these environment variables led to nothing. After some time my team and I figured out how to get debug output from Qt source on our system:
Modifying source code in the qtbase directory under our yocto build (roughly /yocto/poky/build/tmp/work/temp build directory/qtbase
Copying qtbase/plugins/generic/libqevdevmouseplugin.so to my hardware (roughly /usr/lib/qt5/plugins/generic)
Running Qt from the command line
We quickly discovered that the input events coming from mouse0 and mice were basically garbage data. On our system we did set up EVDEV in the kernel, so the mouse input was also tied to the device node event0. When we tried setting the Qt mouse parameter to event0 we started to see debug output that looked like real data.
QT_QPA_GENERIC_PLUGINS=evdevmouse
QT_QPA_EVDEV_MOUSE_PARAMETERS=/dev/input/event0
However, the problem of no-mouse-pointer still remained. After a while we looked back at the Qt Documentation, specifically at the 2nd paragraph listed above. As a last ditch attempt we tried adding in the QT_QPA_FB_HIDECURSOR environment variable...
QT_QPA_GENERIC_PLUGINS=evdevmouse
QT_QPA_EVDEV_MOUSE_PARAMETERS=/dev/input/event0
QT_QPA_FB_HIDECURSOR=0
And...voila! After countless hours of debugging and reading documentation, we finally got a mouse pointer.
I think the main crux of our issue was misinterpreting the Qt Documentation.
The mouse cursor shows up whenever ... QT_QPA_FB_HIDECURSOR (for linuxfb) is not set
By "not set", Qt means explicitly defined as FALSE...not simply "not set" at all.
This solution will work for us, but it does leave at least one thing to be desired. Along the way I stumbled across this thread answer on the Unix StackEx which points to the Kernel documentation of input/input.txt. In section "3.2.2 mousedev" you can see the line:
Each 'mouse' device is assigned to a single mouse or digitizer, except
the last one - 'mice'. This single character device is shared by all
mice and digitizers, and even if none are connected, the device is
present. This is useful for hotplugging USB mice, so that programs
can open the device even when no mice are present.
What this means for us is that while we can use event0 (which goes away when we unplug the mouse) for our mouse input event handling, we won't be able to support hot plugging without making some kernel/Qt-source modifications or figuring out how to get mice working as a Qt mouse input parameter.
So, the question of "why does event0 work and not mouse0/mice" still stands...but for now we've got a solution we can live with.
UPDATE: Now a little bit later we've figured out that udev was not working properly on our system. We added udev to the RDEPENDS in our package group for the Yocto build, and now we can set
QT_QPA_GENERIC_PLUGINS=evdevmouse
and we get a working mouse pointer with hotplug support.
I dont know if this applies to your problem (i dont use QT), but there is a
HAVE_TOUCHSCREEN=1 variable in the machconfig file. It is located normally in your BSP-layer in a recipes-bsp/formfactor/formfactor directory.
Setting this to 1 makes the cursor invisible.
Try setting it to 0
Let me say at outset that I'm using old technology by today's standards! This happens in MFC on Visual Studio 2005 and running under WinXP. (If it 'aint broke.... ;-) )
I have a dialog based app which has a CTabCtrl with two tabs. Each contains a CListCtrl. These work perfectly under normal circumstances. They populate correctly and show and hide as they should. When I first open the application the display selection is correct. If I then minimize the dialog and restore, the CList Ctrl does not show, the tab is blank. It is the only control which has this problem. Another CListCtrl outside of the CTabCtrl does show up correctly. If I then swap tabs and back again, the other tab shows up then the first appears as normal.
This does not happen if I access any other part of the dialog before minimising, it is only when minimising is the absolute first action I take. It also happens with the CListCtrl I have in the other tab if I set this tab to be selected on startup in OnInitDialog where I set up the CTabCtrl.
I have actually solved the effect of this problem by adding into my OnSysCommand(...) the following:
if ((nID & 0xFFF0) == SC_RESTORE)
{
m_ctrlReadList.Invalidate();
}
but it bugs me that I'm adding code to solve a problem which only happens in such odd circumstances. I can't help thinking that there is something I have missed in the setup which is leading to this behaviour. Can anyone offer any explanation as to what is causing it in the first place?
My explanation is based on the facts of what I have found but this has been a learning experience for me so apologies if I get some of it a little confused, I'm still letting it settle in my mind.
Along with this problem I found another which turned out to be relevant. From the nature of the program I'm writing, my CListCtrls needed fixed width headers. Now that turned out to be another thing I couldn't set up! I just needed to prevent the user from grabbing and resizing the header's dividers or double clicking them to autosize, and of course there is functionality in the CListCtrl based on its child CHeaderCtrl to set this up isn't there? Well apparently not. LVS_EX_HEADERDRAGDROP for example isn't the way.
So I explored trying to capture messages which would allow me to myself, and what do you know, I couldn't! I could trap a few but not the ones I needed. I was looking for HDN_BEGINTRACK and HDN_DIVIDERDBLCLICK. (We won't go into the fact that you HAVE TO deal with both A and W versions of those separately!) The CHeaderCtrl is a child of the CListCtrl but it sends its messages back to the CDialog as the CListCtrl's parent. I tried there using both my list's and 0 as ID which headers apparently use. Many of them just plain didn't appear there at all.
So I created my own CListCtrl class inheriting from CListCtrl, overrode OnNotify and they turned up there. I simply prevented CListCtrl::On Notify from being called for those messages and it worked, no resize functionality at all.
I also played with the Z-order too which could have been relevant. I didn't explain earlier but this and another list are on two tabs, exactly aligned over each other. Selecting the tabs HIDEs and SHOWs each of the lists in turn. The other list had no display problems even when I changed the default display in OnInitDialog to show it at start up.
It was under the problematic one. So changing the Z-order in OnInitDialog where I set them up - did nothing! And to rub it in that second list was also unresizeable by default just as I wanted and I couldn't find out why. Their Properties listed exactly the same and there was nowhere in the code where any different aspect was set manually for either, they were effectively theoretically identical, but practically not so. So frustrating!!!
And the upshot of it is that now that the header resize issue is solved my display problem has vanished too! It looks to me and to a few others out there too who report similar symptoms as myself as though the CListCtrl is another one of the slightly flaky ones and needs a little massaging to get the best out of it.
I hope that makes sense to those of you out there who know this control well. I was surprised how simple the solution was, but it also surprised me that the diagnostic process was so difficult. It may of course come down to the ageing system I work within. Nowadays I do this only for fun and the expense of updating VS from 2005 for occasional use is not a high priority. I am sure that some of the symptoms will not show under other build and run environments but it may be worth having the issue and my solution on record somewhere for Google to find for others.
This question is related to: Which API Microsoft Word (Office 2013) is using the paint the screen, which was left unanswered.
I have an Office plug-in (native C++) which is fairly involving with Office. Using hooks I'm modifying the client area of Office application.
Things were working great for me up to Office 2010, hooking WM_NCPAINT, and using GDI method on the Office Window.
But Office 2013 uses Direct2D, DXGI, and DirectWrite. On Windows 7, if I draw something on the screen - it gets erased the next time the cursor blink. On Windows 8 - I don't even have access to the screen.
Currently, my thinking is to hook ID2D1DeviceContext::BeginDraw, ID2D1DeviceContext::EndDraw - and on last function, just before delegating to the real function, I will add my stuff. Another idea is the hook IDXGISwapChain1::Present1. Problem with both these methods that I've hard time identifying what is going on above.
I'm even thinking even a different process (or GDI based window - if possible) on top of the client area I want to decorate, with the majority of the real eastate being 100% transparent. Problem with this solution are issues like keyboard focus and mouse click.
Any tip or suggestion will be appreciated. Any tool (the Spy++ for DirectX) will be appreciated. Does DirectX (specifically Direct2D) has documented extensible or plug-able story, I'm not aware of and could use? Is there anything in Kernel Mode I'm missing?
Hm dont know exactly how that works on Direct Draw, but i hooked once the EndScene function for Direct3D and added some stuff to the scene before the real EndScene could be executed.
I made a little video that shows that:
http://www.youtube.com/watch?v=ZFshqIEaLBc
Here's my advice: don't do that. Any of it. If you find yourself having to hook and hack into the host that deeply, it's a clear sign that you just shouldn't be doing that. Find another way to draw what you need, or change your UI design so that it fits in with what Office already accomodates. Even if you assume it's possible to get this working, every time Office is updated (patch, service pack, new version) you will be at high risk of all your stuff breaking, with no guarantee that you'll be able to get it working again.
I'm working on a I/O verification tool based on Linux in a game project. It is written in C++, and,since using the same I/O module as our game, it's based on OIS 1.2. Thus, though all I need is to print users' inputs on the console, I still need to create a window for OIS.
So here comes my question: How can I create a mapped window while it is still invisible and processes keyboard events?
I can't unmapped the window in that it won't process any keyboard event anymore. I also can't find function for show/hide a window.(maybe I search through a wrong diretion...)
My little tool works fine now except there is a stupid top-level empty window which needs to be focused for processing keyboard events...
Any advise is welcomed.
Thanks!!!
After reading this post: Linux/X11 input library without creating a window,
I realized my problem was that I misunderstood the philosophy of X11. All I need to do is simply pass the root window handle to OIS, and set the x11_grabkeyboard flag as true. The only drawback is maybe I can hardly debug my program with gdb since the keyboard is grabbed...
Though my situation is solved, there is one thing left.
Every article I read said an InputOnly window won't be visible and is capable for handling input events, while my InputOnly window is absolutely visible after mapped...
Maybe it's my Linux, or again, a misunderstanding...