GLUT for Win32 README --------------------- VERSION/INFO: This is GLUT for Win32 version 3.7.6 as of Nov 8th 2001. See the COPYRIGHT section for distribution and copyright notices. Send all bug reports and questions for this version of GLUT to Nate Robins. For more information about GLUT for Win32, see the web page: www.pobox.com/~nate/glut.html or subscribe to the GLUT for Win32 mailing list by sending e-mail to majordomo@perp.com with "subscribe glut" in the body of the message. For general information about GLUT, see the GLUT web page: http://reality.sgi.com/opengl/glut3/glut3.html and be sure to check the GLUT FAQ first for any questions that you may have: http://reality.sgi.com/opengl/glut3/glut-faq.html COMPILING/INSTALLATION: o Precompiled versions of the DLL and import library can be found on the GLUT for Win32 web page mentioned above. o Microsoft Developer Studio 6 workspace and project files have been included in the source code distribution. To build the glut dll: First, open Microsoft Developer Studio. Then, select File -> Open Workspace and find the glut.dsw file in the file dialog and double-click on it. Finally, select Build -> Build glut32.dll. When the build is finished, it will copy: glut32.dll to %WinDir%\System, glut32.lib to $(MSDevDir)\..\..\VC98\lib, and glut.h to $(MSDevDir)\..\..\VC98\include\GL. Additional workspace files have been included in the progs, test and lib directories to build the progs, tests and libs respectively. BORLAND NOTES: From what I understand, Borland supplies a utility that converts Microsoft Visual C++ .libs into Borland compatible files. Therefore, the best method for Borland users is probably to get the precompiled versions of the library and convert the library. To create an import library for Borland from the DLLs, use the following command (from a command prompt): IMPLIB glut32.lib glut32.dll If IMPLIB crashes when called this way, try IMPLIB glut32.lib glut32.def using the glut32.def file in this distribution. FORTRAN NOTES: Bill Mitchell [william.mitchell@nist.gov] has put considerable effort into getting GLUT to work with different compilers for Fortran 90. He indicates that you should copy the f90glut.h file to your $(MSDevDir)\..\..\VC98\include\GL directory. Then, just build GLUT as usual. The Fortran 90 interface, f90gl, can be obtained at http://math.nist.gov/f90gl and contains installation instructions and usage examples. MISC NOTES: o Overlay support is not implemented, nor are there any plans to implement it in the near future. o To customize the windows icon, you can use the resource name GLUT_ICON. For example, create an icon named "glut.ico", and create a file called glut.rc that contains the following: GLUT_ICON ICON glut.ico then compile the glut.rc file with the following: rc /r glut and link the resulting glut.res file into your executable (just like you would an object file). Alternatively, you can simply add the glut.rc file to your project if you are using Microsoft Developer Studio. IMPLEMENTATION DEPENDENT DIFFERENCES: There are a few differences between the Win32 version of GLUT and the X11 version of GLUT. Those are outlined here. Note that MOST of these differences are allowed by the GLUT specification. Bugs and unsupported features are outlined in the UNSUPPORTED/BUGS section. o glutInit: The following command line options have no meaning (and are ignored) in GLUT for Win32: -display, -indirect, -direct, -sync. o glutInitWindowPosition, glutPositionWindow: Win32 has two different coordinate systems for windows. One is in terms of client space and the other is the whole window space (including the decorations). If you glutPositionWindow(0, 0), GLUT for Win32 will place the window CLIENT area at 0, 0. This will cause the window decorations (title bar and left edge) to be OFF-SCREEN, but it gives the user the most flexibility in positioning. HOWEVER, if the user specifies glutInitWindowPosition(0, 0), the window is placed relative to window space at 0, 0. This will cause the window to be opened in the upper left corner with all the decorations showing. This behaviour is acceptable under the current GLUT specification. o glutSetIconTitle, glutSetWindowTitle: There is no separation between Icon title and Window title in Win32. Therefore, setting an icon title in Win32 has no effect. o glutSetCursor: As indicated in the GLUT specification, cursors may be different on different platforms. This is the case in GLUT for Win32. For the most part, the cursors will match the meaning, but not necessarily the shape. Notable exceptions are the GLUT_CURSOR_INFO & GLUT_CURSOR_SPRAY which use the crosshair cursor and the GLUT_CURSOR_CYCLE which uses the 'no' or 'destroy' cursor in Win32. o glutVisibilityFunc: Win32 seems to be unable to determine if a window is fully obscured. Therefore, the visibility of a GLUT window is only reflected by its Iconic, Hidden or Shown state. That is, even if a window is fully obscured, in GLUT for Win32, it is still "visible". o glutEntryFunc: Window Focus is handled differently in Win32 and X. Specifically, the "window manager" in Win32 uses a "click to focus" policy. That is, in order for a window to receive focus, a mouse button must be clicked in it. Likewise, in order for a window to loose focus, a mouse button must be clicked outside the window (or in another window). Therefore, the Enter and Leave notification provided by GLUT may behave differently in the Win32 and in X11 versions. There is a viable workaround for this. A program called "Tweak UI" is provided by Microsoft which can be used to change the focus policy in Win32 to "focus follows mouse". It is available from the Microsoft Web Pages: http://www.microsoft.com/windows/software/PowerToy.htm o glutCopyColormap: GLUT for Win32 always copies the colormap. There is never any sharing of colormaps. This is probably okay, since Win32 merges the logical palette and the physical palette anyway, so even if there are two windows with totally different colors in their colormaps, Win32 will find a (hopefully) good match between them. o glutIdleFunc + menus: The glut idle function will NOT be called when a menu is active. This causes all animation to stop when a menu is active (in general, this is probably okay). Timer functions will still fire, however. If the timer callback draws into the rendering context, the drawing will not show up until after the menu has finished, though. UNSUPPORTED/BUGS: o glutAttachMenu: Win32 only likes to work with left and right mouse buttons. Especially so with popup menus. Therefore, when attaching the menu to the middle mouse button, the LEFT mouse button must be used to select from the menu. o glutSpaceball*, glutButtonBox*, glutTablet*, glutDials*: None of the special input devices are supported at this time. o When resizing or moving a GLUT for Win32 window, no updating is performed. This causes the window to leave "tracks" on the screen when getting bigger or when previously obscured parts are being revealed. I put in a bit of a kludgy workaround for those that absolutely can't have the weird lines. The reshape callback is called multiple times for reshapes. Therefore, in the reshape callback, some drawing can be done. It should probably be limited to a color buffer clear. o The video resizing capabilities of GLUT 3.3+ for X11 is currently unimplemented (this is probably ok, since it really isn't part of the spec until 4.0). I doubt that this will ever be part of GLUT for Win32, since there is no hardware to support it. A hack could simply change the resolution of the desktop. CHANGES/FIXES: (Nov 8, '01) x Released 3.7.6 (Nov 8, '01) x Changed fullscreen mode from TOPMOST back to simply TOP, since (it turns out) many people use windows atop a GLUT window. (Nov 8, '01) x Added code to prevent CPU spiking when no idle function is registered. Otherwise, if an idle function is registered, spike CPU so that the idle function gets all the attention it needs and if this is a problem on the program side, the user can stick a sleep() in their idle function. I believe that this strikes the best balance betweeen GLUT being fast, and also being "nice" to other processes. Thanks to James Wright for reporting this bug. (Nov 8, '01) x Fixed bug in motion callback handler which wasn't setting the current window, so multiple window apps (e.g., any GLUI app) wouldn't get the callback correctly. (Oct 4, '01) x Fixed bug in glutEnterGameMode() that caused new windows to not be in "fullscreen" mode, so they got window decorations. (Oct 4, '01) x Fixed bug in glutEnterGameMode() that caused new windows to not be in "fullscreen" mode, so they got window decorations. (Oct 3, '01) x Fixed bug in getVisualInfoFromString(): visuals not reloaded on display mode change. Reload visuals each time they are queried. This fixes a problem with Win32 because the list of availabe Visuals (Pixelformats) changes after a change in displaymode. The problem occurs when switching to gamemode and back. Thanks to Michael Wimmer for pointing this out & providing the fix. (Oct 3, '01) x Fixed bug in XGetVisualInfo(): pixelformats enumerated incorrectly. Passing 0 as a pixelformat index to DescribePixelFormat gives unpredictible results (e.g., this fails on the Voodoo opengl32.dll and always reports 0 as the last available pixelformat index). Thanks to Michael Wimmer for pointing this out & providing the fix. (Oct 3, '01) x Fixed bug in glXGetConfig(): pixelformats enumerated incorrectly. The test was OpenGL support OR draw to window, but should be AND. Thanks to Michael Wimmer for pointing this out & providing the fix. (Sep 28, '01) x Fixed glutChangeToSubMenu()/glutChangeToMenuEntry() bug where if you went back and forth between a submenu and a plain entry, the submenu wouldn't be updated properly. (Sep 28, '01) x glutSetIconTitle() is now a nop. (Sep 28, '01) x glutFullScreen() now sets the window as TOPMOST, therefore, the window will always be on top (this essentially disables alt-tabbing). (Sep 28, '01) x The key repeat ignore flag is now honored correctly. (Sep 28, '01) x Key presses are now reported more accurately and fully, in particular, modified up events (i.e., SHIFT-2) are now reported correctly. (Sep 28, '01) x Subwindows nested arbitrarily deep get their keyboard callbacks correctly now. (Sep 28, '01) x Major rewrite of the window procedure code to clean it up and make way for other bug fixes. (Sep 23, '01) x Fixed noof example program to use RAND_MAX instead of assumed max of 2147483647.0. (Now it looks _much_ better!) (Sep 22, '01) x Fixed sunlight example program. globe.raw data file was corrupt, added a new one. (Sep 22, '01) x Fixed zcomposite example program to print message if overlay support is not found (instead of crashing). (Jan 22, '01) x Fixed malloc(0) bug in Win32 version of XGetVisualInfo. Thanks to Kekoa Proudfoot for bringing this to my attention. (Dec 12, '00) x Added data files for the advanced & advanced97 programs. (Dec 12, '00) x Added Developer Studio 6 project and workspace files for pretty much everything (the stuff left out was usually unix specific). (Dec 7, '00) x Fixed several compilation problems & corrupt files. Thanks to Alexander Stohr for bringing these to my attention and providing detailed fixes. (Dec 6, '00) x Fixed compiler support for lcc. Thanks to Gordon for bringing this to my attention and debugging fixes. (Nov 8, '00) x Fixed submenu problem (sometimes the menu callback was not called for valid items). Thanks to Michael Keeley. (Oct 16, '00) x Corrected corrupt duck.iv file. Thanks to Jon Willeke for finding this problem. (Sept 27, '00) x Fixed bug in processWorkList that could cause a hang. Thanks to Bill Volz & Daniel Azuma. (Sept 26, '00) x Added mui DLL project file (thanks to DMWeldy@ugsolutions.com). (Sept 9, '00) x Fixed Delete key bug (crash when no keyboard callback was registered, but a special key callback was). Thanks to Kent Bowling (kent_bowling@hotmail.com) for finding this bug. (May 18, '00) x Fixed subwindow keyboard callbacks. (May 22, '97) o Menus don't work under Windows 95 x Fixed! Added a unique identifier to each menu item, and a search function to grab a menu item given the unique identifier. (May 21, '97) o A few minor bug fixes here and there. x Thanks to Bruce Silberman and Chris Vale for their help with this. We now have a DLL! (Apr 25, '97) o DLL version of the library is coming (as soon as I figure out how to do it -- if you know, let me know). x Thanks to Bruce Silberman and Chris Vale for their help with this. We now have a DLL! (Apr 24, '97) x Added returns to KEY_DOWN etc messages so that the F10 key doesn't toggle the system menu anymore. (Apr 7, '97) o Palette is incorrect for modes other than TrueColor. x Fixed this by forcing a default palette in modes that aren't Truecolor in order to 'simulate' it. The applications program shouldn't have to do this IMHO, but I guess we can't argue with Microsoft (well, we can, but what good will it do?). (Apr 2, '97) x Added glut.ide file for Borland users. (Apr 2, '97) x Fixed a bug in the WM_QUERYNEWPALETTE message. Wasn't checking for a null colormap, then de-ref'd it. Oops. (Mar 13, '97) o glutTimerFunc: Currently, GLUT for Win32 programs busy waits when there is an outstanding timer event (i.e., there is no select() call). I haven't found this to be a problem, but I plan to fix it just because I can't bear the thought of a busy wait. x Added a timer event and a wait in the main loop. This fixes the CPU spike. (Mar 11, '97) x Fixed subwindow visibility. The visibility callback of subwindows wasn't being called, now it is. (Mar 11, '97) o glutGetHDC, glutGetHWND: In order to support additional dialog boxes, wgl fonts, and a host of other Win32 dependent structures, two functions have been added that operate on the current window in GLUT. The first (glutGetHDC) returns a handle to the current windows device context. The second (glutGetHWND) returns handle to the current window. x Took these out to preserve GLUT portability. (Mar 11, '97) x Fixed the glutWarpPointer() coordinates. Were relative to the screen, now relative to window (client area) origin (which is what they're supposed to be). (Mar 11, '97) o glutCreateMenu, glutIdleFunc: Menu's are modal in Win32. That is, they don't allow any messages to be processed while they are up. Therefore, if an idle function exists, it will not be called while processing a menu. x Fixed! I've put in a timer function that fires every millisecond while a menu is up. The timer function handles idle and timer events only (which should be the only functions that are firing when a menu is up anyway). (Mar 7 '97) x Fixed minor bugs tracked down by the example programs. (Mar 6, '97) x Merged 3.3 GLUT for X11 into 3.2 GLUT for Win32. New code structure allows for EASY merging! o In Win32, the parent gets the right to set the cursor of any of its children. Therefore, a child windows cursor will 'blink' between its cursor and its parent. x Fixed this by checking whether the cursor is in a child window or not. (Feb 28 '97) o On initial bringup apps are getting 2 display callbacks. x Fixed by the Fev 28 re-write. o Some multiple window (not subwindow) functionality is messed up. See the sphere.exe program. x Fixed by the Feb 28 re-write. o GLUT for Win32 supports color index mode ONLY in a paletted display mode (i.e., 256 or 16 color mode). x Fixed this in the re-write. If you can get a color index visual (pixel format) you can use color index mode. (Feb 28 '97) o Quite a few bugs (and incompatibilities) were being caused by the structure that I used in the previous port of GLUT. Therefore I decided that it would be best to "get back to the roots". I re-implemented most of glut trying to stick with the structure layed out by Mark. The result is a much more stable version that passes ALL (!) (except overlay) the tests provided by Mark. In addition, this new structure will allow future enhancements by Mark to be integrated much more quickly into the Win32 version. Also, I'm now ordering the bugs in reverse, so that the most recently fixed appear at the top of the list. (9/8/96) o Changed the glutGetModifiers code to produce an error if not called in the core input callbacks. (9/11/96) o If the alt key is pressed with more than one other modifier key it acts as if it is stuck -- it stays selected until pressed and released again. x Fixed. (9/12/96) o When a submenu is attached to a menu, sometimes a GPF occurs. Fixed. Needed to set the submenu before referencing it's members. o Kenny: Also, one little problem, I attached the menu to the right-button, but when the left-button is pressed I detach it to give the right-button new meaning; if I pop-up the menu and I don't want to select anything, like most users, I click off of the menu to make it disappear. When I do this, I get a GLUT error and the program terminates because I am altering the menu attachment from within the button press while the menu is active. x Fixed. Needed to finish the menu when the user presses the button, not just when a button is released. o GLUT for Win32 emulates a middle mouse button by checking if both mouse buttons are down. This causes a lot of problems with the menu and other multiple mouse button things. x Fixed. No more middle mouse button emulation. Perhaps it would be a good idea to emulate the middle mouse button (if not present) with a key? (9/15/96) o Added code to accept a user defined icon. If no icon is provided, a default icon is loaded. (9/19/96) o Shane: Command line options seem to be screwed up. (9/13) x Fixed. The geometry command line was broken, and so was the gldebug command line. o Fixed a bug in the default glut reshape. It was looking for the parent of the current window and GPF'ing if there wasn't a parent. Put in a check for a parent, and if none is there, use the child. o Idle function sucks up all processor cycles. (9/8/96) x I don't know if this is avoidable. If you have a tight rendering loop, it may be that the processor time is going to be sucked up no matter what. You can add a sleep() to the end of your render loop if you would like to yeild to other processes and you don't care too much about the speed of your rendering loop. If you have Hardware that supports OpenGL (like a 3Dpro card, or GLint card) then this should be less of a problem, since it won't be rendering in software. (9/11/96) o If a window is fully obscured by another window, the visibility callback is NOT called. As far as I can tell, this is a limitation of the Win32 api, but a workaround is being searched for. (9/8/96) x Limitation of the Win32 API o Fixed the entry functions. They only work if the keyboard focus changes. Therefore, in most Win32 systems, the mouse must be pressed outside of the window to get a GLUT_LEFT message and then pressed inside the window for a GLUT_ENTERED message. o Alt modifier key doesn't work with keyboard callback. (9/8/96) x Probably okay, because the glut spec says that these keys can be intercepted by the system (which the alt key is...) (9/11/96) (11/17/96) o glutRemoveMenuItem() not working properly. x Thanks to Gary (grc@maple.civeng.rutgers.edu) for the fix to this one. o Timer functions are messed up. x Thanks to Joseph Galbraith for the fix to this one. (12/9/96) o One (minor) difference came up between the X version of glut and the nt one which you should know about. It is not a new problem, and it concerns co-ords returned to the pointer callbacks. (glutMotionFunc, glutMouseFunc) Under X, you get co-ords in the range 0 +/- 2^15, under NT you get 0..2^16. This is only really a problem when moving above or to the left of the window. eg dragging one pixel ABOVE the window will give :- under x11 : y = -1 under nt : y = 2^16 -1 x Put in fix provided by Shane Clauson. (12/17/96) o Idle functions not working properly for multiple windows. x Fixed this by posting an idle message to every window in the window list when idle. (12/18/96) o glutSetCursor() was misbehaving (lthomas@cco.caltech.edu). x Win32 requires that the hCursor member of the window class be set to NULL when the class is registered or whenever the mouse is moved, the original cursor is replaced (go figure!). Now sets the cursor whenever a WM_MOUSEMOVE message is received, because the WM_SETCURSOR event resets the cursor even when in the decoration area. o Geometry is not being handled quite right. The numbers don't take into account the window decorations. That is, if I say make a window 100x100, then the WHOLE window (not just the client area) is 100x100. Therefore, the client (opengl) area is smaller than 100x100. (9/8/96) x Fixed. Added code to subtract the decoration size on glutGet() and add the decoration size on glutReshapeWindow(). o Multiple glutPostRedisplay() calls are NOT being combined. To get round the "coalesce" problem on glutPostRedisplay, the easiest solution is to roll-your-own coalesce by keeping a global "dirty" flag in the app (eg replace all calls to glutPostRedisplay with image_dirty=TRUE;), and to handle image_dirty with a single glutPostRedisplay in the idle callback when required. (erk - but increases performance for my particular app (a rendering engine on the end of a pipleine with a stream of graphics updates) by a couple of orders of magnitude ! ) (9/8/96) x Added code to coalesce redisplays. Every idle cycle, a check is made to see which windows need redisplay, if they need it, a redisplay is posted. The glutPostRedisplay() call is just a stub that sets a flag. THANKS: Special thanks to the following people for extensive testing, suggestions, fixes and help: Alexander Stohr Shane Clauson Kenny Hoff Richard Readings Paul McQuesten Philip Winston JaeWoo Ahn Joseph Galbraith Paula Higgins Sam Fortin Chris Vale Bill Mitchell and of course, the original author of GLUT: Mark Kilgard. and many others... COPYRIGHT: The OpenGL Utility Toolkit distribution for Win32 (Windows NT & Windows 95) contains source code modified from the original source code for GLUT version 3.3 which was developed by Mark J. Kilgard. The original source code for GLUT is Copyright 1997 by Mark J. Kilgard. GLUT for Win32 is Copyright 1997 by Nate Robins and is not in the public domain, but it is freely distributable without licensing fees. It is provided without guarantee or warrantee expressed or implied. It was ported with the permission of Mark J. Kilgard by Nate Robins. THIS SOURCE CODE IS PROVIDED "AS IS" WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OR MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. OpenGL (R) is a registered trademark of Silicon Graphics, Inc.