RB3D is Broken (Was: Problem with Quesa Declares on XP and RB2006r4 and up) (REALbasic games Mailinglist archive)
Back to the thread list
Previous thread: Trimesh.Handle feature request
Next thread: Frames around several tiles
| RB3D is Broken (Was: Problem with Quesa Declares on XP and RB2006r4 and up) |
| Date: 30.01.07 22:43 (Tue, 30 Jan 2007 16:43:55 -0500) |
| From: Frank Condello |
|
On 26-Jan-07, at 5:07 PM, Frank Condello wrote:
> I was recently made aware that my Quesa Wrappers failed to work on > Windows XP when compiled with RB2006r4 or RB2007r1 but worked fine > with 2006r3. A little digging revealed that creating an Object3D no > longer initializes Quesa (an implementation detailed my wrappers > rely on). Errrm, this is an implementation detail RB3D relies on! RB3D just plain doesn't work on Windows in my testing: <http:// www.realsoftware.com/feedback/viewreport.php?reportidowdxvgr> The "not a bug" evaluation was hasty at best... :/ Frank. <http://developer.chaoticbox.com/> <http://macgameblog.com/> _______________________________________________ Unsubscribe or switch delivery mode: <http://www.realsoftware.com/support/listmanager/> Search the archives of this list here: <http://support.realsoftware.com/listarchives/lists.html> |
| Re: RB3D is Broken (Was: Problem with Quesa Declares on XP and RB2006r4 and up) |
| Date: 31.01.07 19:08 (Wed, 31 Jan 2007 19:08:06 +0100) |
| From: Franco Vaccari |
|
I tried your Primitives example from the Quesa wrappers, and got many
failed assertions like this, probably one per primitive: Runtime Error 4: Failed Assertion loaderX86.cpp: 282 Failure Condition: functionEntry Could not resolve function 'Q3Error_Register' in Quesa Compiled with 2007r1 Mac IDE, run with Win XP under Parallels on a Intel iMac But the RB3D based program I'm working on now runs under Windows (without any Quesa declare), so the statement "RB3D just plain doesn't work on Windows" is not true here, as well as the sentence "Install Quesa an run any RB3D project - you get a blank display (at least I do)" that I see in the feedback report. I have to add that I'm unable to track a crash that later occurs in my own app (only in Win32, not in Mac OS X): after closing all windows I get a crash if I open a new one, with no way of falling back to the remote debugger. Possibly a side effect of something dirty related to your findings... HTH Franco On Jan 30, 2007, at 10:43 PM, Frank Condello wrote: > On 26-Jan-07, at 5:07 PM, Frank Condello wrote: > >> I was recently made aware that my Quesa Wrappers failed to work on >> Windows XP when compiled with RB2006r4 or RB2007r1 but worked fine >> with 2006r3. A little digging revealed that creating an Object3D >> no longer initializes Quesa (an implementation detailed my >> wrappers rely on). > > Errrm, this is an implementation detail RB3D relies on! RB3D just > plain doesn't work on Windows in my testing: <http:// > www.realsoftware.com/feedback/viewreport.php?reportid{wdxvgr> The > "not a bug" evaluation was hasty at best... :/ > > Frank. > <http://developer.chaoticbox.com/> > <http://macgameblog.com/> _______________________________________________ Unsubscribe or switch delivery mode: <http://www.realsoftware.com/support/listmanager/> Search the archives of this list here: <http://support.realsoftware.com/listarchives/lists.html> |
| Re: RB3D is Broken (Was: Problem with Quesa Declares on XP and RB2006r4 and up) |
| Date: 31.01.07 20:04 (Wed, 31 Jan 2007 14:04:55 -0500) |
| From: Frank Condello |
|
On 31-Jan-07, at 1:08 PM, Franco Vaccari wrote:
> I tried your Primitives example from the Quesa wrappers, and got > many failed assertions like this, probably one per primitive: > > Runtime Error 4: Failed Assertion > > loaderX86.cpp: 282 > Failure Condition: functionEntry > Could not resolve function 'Q3Error_Register' in Quesa That looks like an entirely different problem... a functionEntry failure means it can't find Quesa.Framework, and I believe that's RB3D firing the assertion. The hard declares in the Quesa Wrappers would fail on launch with a system-supplied error dialog. > Compiled with 2007r1 Mac IDE, run with Win XP under Parallels on a > Intel iMac > > But the RB3D based program I'm working on now runs under Windows > (without any Quesa declare), so the statement "RB3D just plain > doesn't work on Windows" is not true here, as well as the sentence > "Install Quesa an run any RB3D project - you get a blank display > (at least I do)" that I see in the feedback report. Well I did qualify that statement with "at least I do" ;) All indications are, that on my XP box, Quesa is never initialized by RB3D. It's possible you're using some RB3D object that does initialize Quesa at some point, whereas my test project never does. Without the RB3D plugin source code in front of me I can't tell anyone what's really wrong, I can only make an educated guess... > I have to add that I'm unable to track a crash that later occurs in > my own app (only in Win32, not in Mac OS X): after closing all > windows I get a crash if I open a new one, with no way of falling > back to the remote debugger. Possibly a side effect of something > dirty related to your findings... It's possible that some RB3D objects initialize Quesa, but others shut it down pre-maturely. From what I understand RB3D basically ref- counts Quesa initializations (so you don't have to start/stop it manually), and if those locks/unlocks get out of whack really bad things can happen. Frank. <http://developer.chaoticbox.com/> <http://macgameblog.com/> _______________________________________________ Unsubscribe or switch delivery mode: <http://www.realsoftware.com/support/listmanager/> Search the archives of this list here: <http://support.realsoftware.com/listarchives/lists.html> |
| Re: RB3D is Broken (Was: Problem with Quesa Declares on XP and RB2006r4 and up) |
| Date: 31.01.07 22:00 (Wed, 31 Jan 2007 15:00:18 -0600) |
| From: David Grogono |
|
Hi Frank -
As another data point I have many RB3D apps that are all working fine on Windows. I'm guessing that your projects all use Declares and mine (and Franco's) do not. I believe there was a change not that long ago where Quesa is built in to the compiled Windows app as it is on Linux. I thought this change occurred before 2006r4 but perhaps that was the first version with the change. Could this explain what you are seeing? - David On Jan 31, 2007, at 1:04 PM, Frank Condello wrote: > On 31-Jan-07, at 1:08 PM, Franco Vaccari wrote: > >> I tried your Primitives example from the Quesa wrappers, and got >> many failed assertions like this, probably one per primitive: >> >> Runtime Error 4: Failed Assertion >> >> loaderX86.cpp: 282 >> Failure Condition: functionEntry >> Could not resolve function 'Q3Error_Register' in Quesa > > That looks like an entirely different problem... a functionEntry > failure means it can't find Quesa.Framework, and I believe that's > RB3D firing the assertion. The hard declares in the Quesa Wrappers > would fail on launch with a system-supplied error dialog. > >> Compiled with 2007r1 Mac IDE, run with Win XP under Parallels on a >> Intel iMac >> >> But the RB3D based program I'm working on now runs under Windows >> (without any Quesa declare), so the statement "RB3D just plain >> doesn't work on Windows" is not true here, as well as the sentence >> "Install Quesa an run any RB3D project - you get a blank display >> (at least I do)" that I see in the feedback report. > > Well I did qualify that statement with "at least I do" ;) All > indications are, that on my XP box, Quesa is never initialized by > RB3D. It's possible you're using some RB3D object that does > initialize Quesa at some point, whereas my test project never does. > Without the RB3D plugin source code in front of me I can't tell > anyone what's really wrong, I can only make an educated guess... > >> I have to add that I'm unable to track a crash that later occurs >> in my own app (only in Win32, not in Mac OS X): after closing all >> windows I get a crash if I open a new one, with no way of falling >> back to the remote debugger. Possibly a side effect of something >> dirty related to your findings... > > It's possible that some RB3D objects initialize Quesa, but others > shut it down pre-maturely. From what I understand RB3D basically > ref-counts Quesa initializations (so you don't have to start/stop > it manually), and if those locks/unlocks get out of whack really > bad things can happen. > > Frank. > <http://developer.chaoticbox.com/> > <http://macgameblog.com/> > _______________________________________________ > Unsubscribe or switch delivery mode: > <http://www.realsoftware.com/support/listmanager/> > Search the archives of this list here: > <http://support.realsoftware.com/listarchives/lists.html> _______________________________________________ Unsubscribe or switch delivery mode: <http://www.realsoftware.com/support/listmanager/> Search the archives of this list here: <http://support.realsoftware.com/listarchives/lists.html> |
| Re: RB3D is Broken (Was: Problem with Quesa Declares on XP and RB2006r4 and up) |
| Date: 31.01.07 22:52 (Wed, 31 Jan 2007 16:52:47 -0500) |
| From: Frank Condello |
|
On 31-Jan-07, at 4:00 PM, David Grogono wrote:
> As another data point I have many RB3D apps that are all working > fine on Windows. I'm guessing that your projects all use Declares > and mine (and Franco's) do not. A plain RB3D project comes up blank for me - no declares - but I'll do some more testing when I have time (i.e. in a couple weeks). The project in question was a lighting test so maybe Quesa's point lights are broken on Windows and the scene was black due to that... > I believe there was a change not that long ago where Quesa is built > in to the compiled Windows app as it is on Linux. I thought this > change occurred before 2006r4 but perhaps that was the first > version with the change. Could this explain what you are seeing? Declares work with 2006r3, but not in 2006r4 or later*. I don't recall a release note about this but I know several people who'll be quite upset if declares stopped working on Windows... :/ * If you force-initialize Quesa with a declare things seem to work OK - not that I tested extensively - this suggests RB3D is linking with the external DLL, 'cause IIRC, you can't link to two separate Quesa libraries and not crash horribly - though perhaps Windows expresses this as a slow death instead of the immediate crash you'd see on OS X. If Quesa is statically linked the evaluation to bug #nbwdxvgr would still be bogus however. Frank. <http://developer.chaoticbox.com/> <http://macgameblog.com/> _______________________________________________ Unsubscribe or switch delivery mode: <http://www.realsoftware.com/support/listmanager/> Search the archives of this list here: <http://support.realsoftware.com/listarchives/lists.html> |