Re: [Flightgear-devel] strange screen
On Tue, 2012-09-04 at 22:14 +0100, Alasdair wrote: On Tue, 2012-09-04 at 21:58 +0100, Alasdair wrote: On Tue, 2012-09-04 at 22:34 +0200, Frederic Bouvier wrote: If you start FG with --log-level=debug --log-class=view you'll be able to see messages like : FGEventHandler::handle: RESIZE event 763 x 1338, resizable: 1 FGRenderer::resize: new size 1338 x 763 What are yours ? Regards, -Fred I see FGRenderer::resize: new size 1904 x 993 but no reference to FGEventHandler::? I already has a look at FGRenderer::resize, but the method doesn't seem to do anything, except issue the above message at debug level. Keep going, Fred! I have every confidence you can crack this problem. It's the FGEventHandler::handle that actually do the resize. So if you don't see it it explains what you see. BTW: What is your OSG version ? Regards, -Fred 3.0.1 I always used my own build, but also tried debian's deb (version 3.0.1-3 +b2) Alasdair I see : FGEventHandler::handle: RESIZE event 1002 x 1920, resizable: 1 FGRenderer::resize: new size 1920 x 1002 but _only after I click the maximize widget_ Screen reverts to small view if I click any menu item Alasdair Mind you, I don't see FGEventHandler::handle in the log for checkout prior to caeebdbd935015fd7b2bb3071e5f09b1b7815415 either, where the problem is not evident. Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
I have narrowed this problem down further. Comment out: instrumentation n=0 pathAircraft/c172p/Systems/instrumentation.xml/path /instrumentation in c172p-set.xml Similarly instrumentation pathAircraft/b1900d/Systems/instrumentation.xml/path /instrumentation -- in b1900d-set.xml and in the case of the SenecaII instrumentation pathAircraft/SenecaII/Systems/SenecaII-instruments.xml/path /instrumentation in SenecaII-base.xml The Calalyst view port problem disappears and the problem planes behave beautifully. Can someone with this problem confirm? Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On Wed, 5 Sep 2012, Alasdair wrote: I have narrowed this problem down further. Comment out: instrumentation n=0 pathAircraft/c172p/Systems/instrumentation.xml/path /instrumentation in c172p-set.xml Similarly instrumentation pathAircraft/b1900d/Systems/instrumentation.xml/path /instrumentation -- in b1900d-set.xml and in the case of the SenecaII instrumentation pathAircraft/SenecaII/Systems/SenecaII-instruments.xml/path /instrumentation in SenecaII-base.xml The Calalyst view port problem disappears and the problem planes behave beautifully. Can someone with this problem confirm? That would make the aircraft use the generic instrumentation file that includes the wxradar instrument (which usually or, maybe, always masks the problem). You could also try adding radar namewxradar/name number0/number /radar in the aircraft's own instrumentation file. This, however, is just a work-around, not a solution. Cheers, Anders -- --- Anders Gidenstam WWW: http://gitorious.org/anders-hangar http://www.gidenstam.org/FlightGear/ -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On Wed, 2012-09-05 at 15:47 +0200, Anders Gidenstam wrote: On Wed, 5 Sep 2012, Alasdair wrote: I have narrowed this problem down further. Comment out: instrumentation n=0 pathAircraft/c172p/Systems/instrumentation.xml/path /instrumentation in c172p-set.xml Similarly instrumentation pathAircraft/b1900d/Systems/instrumentation.xml/path /instrumentation -- in b1900d-set.xml and in the case of the SenecaII instrumentation pathAircraft/SenecaII/Systems/SenecaII-instruments.xml/path /instrumentation in SenecaII-base.xml The Calalyst view port problem disappears and the problem planes behave beautifully. Can someone with this problem confirm? That would make the aircraft use the generic instrumentation file that includes the wxradar instrument (which usually or, maybe, always masks the problem). You could also try adding radar namewxradar/name number0/number /radar in the aircraft's own instrumentation file. This, however, is just a work-around, not a solution. Cheers, Anders Sort of back where we started really, but I noticed (as I guess you know) that it is not the contents of the instrumentation file, but its mere inclusion that causes the problem. Ie replace the instrumentation file with an empty(ish) xml file and the problem persists (and the darn instruments don't work). I will try your suggested fix, rather than trashing the instrumentation files. Thanks. Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On Wed, 2012-09-05 at 15:06 +0100, Alasdair wrote: On Wed, 2012-09-05 at 15:47 +0200, Anders Gidenstam wrote: On Wed, 5 Sep 2012, Alasdair wrote: I have narrowed this problem down further. Comment out: instrumentation n=0 pathAircraft/c172p/Systems/instrumentation.xml/path /instrumentation in c172p-set.xml Similarly instrumentation pathAircraft/b1900d/Systems/instrumentation.xml/path /instrumentation -- in b1900d-set.xml and in the case of the SenecaII instrumentation pathAircraft/SenecaII/Systems/SenecaII-instruments.xml/path /instrumentation in SenecaII-base.xml The Calalyst view port problem disappears and the problem planes behave beautifully. Can someone with this problem confirm? That would make the aircraft use the generic instrumentation file that includes the wxradar instrument (which usually or, maybe, always masks the problem). You could also try adding radar namewxradar/name number0/number /radar in the aircraft's own instrumentation file. This, however, is just a work-around, not a solution. Cheers, Anders Sort of back where we started really, but I noticed (as I guess you know) that it is not the contents of the instrumentation file, but its mere inclusion that causes the problem. Ie replace the instrumentation file with an empty(ish) xml file and the problem persists (and the darn instruments don't work). I will try your suggested fix, rather than trashing the instrumentation files. Thanks. Alasdair Clearly, as has been widely suggested, the inclusion of the radar stanza in the instrumentation file(s) does indeed mask this infuriating problem. It may be time to take this discussion off-list? I have no idea as to how to even _start_ to debug this problem, but I have the time, the resources, the will, and maybe even the skill to have a go. What I lack is the knowledge. Anyone wanting to contact me (maybe off-list) do not hesitate. We have some of the best coders in the world. Lend me a hand to get started, and I will give it my best shot. Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On Mon, 2012-09-03 at 21:32 +0100, Alasdair wrote: On Mon, 2012-09-03 at 09:39 -0600, syd adams wrote: Im running Ubuntu , still get that problem.I've added the radar to my intrumentation file in my affected aircraft , and the problem is gone.Strange. Syd On Mon, Sep 3, 2012 at 2:23 AM, Alasdair ali...@btinternet.com wrote: On Mon, 2012-09-03 at 01:06 +0200, ThorstenB wrote: On Sun, Sep 2, 2012 at 11:31 AM, Alasdair ali...@btinternet.com wrote: After git pull (sg,fg,fgdata) a couple of days ago, I was presented with a strange screen which I had not seen before. My regular was shown as a sub-screen in the bottom-left corner. Clicking the maximise widget restores full screen, but any click on a menu item returns me to this stupid screen. Is this infuriating behaviour deliberate, or has my graphics driver gone walk-about? https://dl.dropbox.com/u/15936159/fg_quarter-screen.png Unfortunately a long standing issue. This affects ATI graphics cards when running newer drivers. It can be fixed by reverting to an older ATI driver (see link below). To me it looked like a genuine OSG or ATI driver bug (the size of the graphics view port doesn't properly resize to the size of the window), but it's likely triggered by something special we're doing in FG, since other applications don't seem to have this issue. Also, it's timing-/sequence dependant: it only affects certain aircraft (aircraft using wxradar, which changes/delays the OSG init sequence on startup, are less likely to have the issue) - and for some people it dis- and reappears with certain FG builds. It would be really good if someone seeing the issue locally could properly debug this - even if it means digging into OSG - since a larger number of people are affected. I tried remote debugging this a while ago, but gave up after being unable to narrow it down inside FG. Maybe someone else has better luck. For more see here: http://code.google.com/p/flightgear-bugs/issues/detail?id=385 cheers, Thorsten Thanks very much for the link, Thorsten, it is a good start. But strange that this problem, should just appear without any driver change. But I am using Debian SID, and update my system daily, so it could be OS-related. I will install UBUNTU and try the same build under that OS. regards, Alasdair OK I tried building the same git version on Ubuntu (different gcc version) Same problem. Tried a complete fresh clone (sg/fg/fgdata) Same problem. Downloaded and built 2.8.0 and built with same script : NO problem. So this is a git problem and fairly recent too. 0- Most aircraft I have tested behave properly, but the ones that have this strange viewport problem include : c172p, b1900d, SenecaII. The viewport changes immediately after the splash screen displays initializing graphics engine. Maybe someone can figure out what these have in common that the 707, 737-300, 777-200ER, A-10, A300, aerostar700, alphajet, b29, bf109g, Bravo, c310, Citation*, CRJ700ER, Cub and pa24-250-CIIB do not. I cannot therefore attribute this problem to the Catalyst driver. My git pull/build of 12 Aug did not display this problem. That of 1 Sep did. So I reckon it must have been a commit made between these dates. Alasdair Further research shows that for the b1900d SenecaII, this problem goes way back to 2.6.0. It has only become apparent (in my case) for the c172p in recent git pulls. I will attempt to find exactly when, but I am having to learn git as I go :-( Since the c172p is my aircraft of choice, the problem has gone unnoticed up till now. Thanks to all who are helping me resolve this issue. Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On 4 Sep 2012, at 11:43, Alasdair wrote: Further research shows that for the b1900d SenecaII, this problem goes way back to 2.6.0. It has only become apparent (in my case) for the c172p in recent git pulls. I will attempt to find exactly when, but I am having to learn git as I go :-( Since the c172p is my aircraft of choice, the problem has gone unnoticed up till now. I would be very cautious in your testing - from what Thorsten's already indicated, you are /not/ looking at a stable issue. There's certainly code-paths that bias it (wxradar apparently, and model complexity / texture complexity in general), but changing OS / driver / FG version and having the problem go away or re-appear doesn't conclusively indicate that a specific change was made in those things, that is causal for the bug. In other words, changing your configuration (aircraft, rendering settings, drivers, who knows) may push you back-and-forth over the line where the bug occurs, but that doesn't tell us much about the real cause. Of course, testing is still worthwhile - it would be interesting to know if anyone saw the issue with FG 2.4 - but again so much has changed in terms of shaders complexity, that I fear this again might not prove anything. James -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On Tue, 2012-09-04 at 12:07 +0100, James Turner wrote: On 4 Sep 2012, at 11:43, Alasdair wrote: Further research shows that for the b1900d SenecaII, this problem goes way back to 2.6.0. It has only become apparent (in my case) for the c172p in recent git pulls. I will attempt to find exactly when, but I am having to learn git as I go :-( Since the c172p is my aircraft of choice, the problem has gone unnoticed up till now. I would be very cautious in your testing - from what Thorsten's already indicated, you are /not/ looking at a stable issue. There's certainly code-paths that bias it (wxradar apparently, and model complexity / texture complexity in general), but changing OS / driver / FG version and having the problem go away or re-appear doesn't conclusively indicate that a specific change was made in those things, that is causal for the bug. In other words, changing your configuration (aircraft, rendering settings, drivers, who knows) may push you back-and-forth over the line where the bug occurs, but that doesn't tell us much about the real cause. Of course, testing is still worthwhile - it would be interesting to know if anyone saw the issue with FG 2.4 - but again so much has changed in terms of shaders complexity, that I fear this again might not prove anything. James Thanks for the words of wisdom, James, but I am still curious to determine when the problem began for me on the c172p. I don't think Fred's suggestion for git bisect will work because I have to bisect fg, sg and data and keep them in sync, so I reckon I will have to do the binary chop manually. This is really a labour of love curiosity, cos all is ok on another more powerful m/c with nvidia graphics + 4 monitors. regards, Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On Tue, 4 Sep 2012, Alasdair wrote: Thanks for the words of wisdom, James, but I am still curious to determine when the problem began for me on the c172p. I don't think Fred's suggestion for git bisect will work because I have to bisect fg, sg and data and keep them in sync, so I reckon I will have to do the binary chop manually. This is really a labour of love curiosity, cos all is ok on another more powerful m/c with nvidia graphics + 4 monitors. Try checking out the c172p as of the commit before caeebdbd935015fd7b2bb3071e5f09b1b7815415 E.g. git checkout caeebdbd935015fd7b2bb3071e5f09b1b7815415^ -- Aircraft/c172p That commit, from Aug 23, switches the c172p to use its own instrumentation file rather than using the generic one. While the generic one contains the wxradar instrument, the specific (naturally) doesn't. IIRC enabling the wxradar instrument is a well/long-known work-around for this problem on ATI hardware? Also try changing the default window size settings in preferences.xml - to see if the frame buffer for some reason reverts to the default/startup size after the windows is resized to fullscreen. Cheers, Anders -- --- Anders Gidenstam WWW: http://gitorious.org/anders-hangar http://www.gidenstam.org/FlightGear/ -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On Tue, 2012-09-04 at 15:00 +0200, Anders Gidenstam wrote: Try checking out the c172p as of the commit before caeebdbd935015fd7b2bb3071e5f09b1b7815415 E.g. git checkout caeebdbd935015fd7b2bb3071e5f09b1b7815415^ -- Aircraft/c172p That commit, from Aug 23, switches the c172p to use its own instrumentation file rather than using the generic one. While the generic one contains the wxradar instrument, the specific (naturally) doesn't. IIRC enabling the wxradar instrument is a well/long-known work-around for this problem on ATI hardware? Also try changing the default window size settings in preferences.xml - to see if the frame buffer for some reason reverts to the default/startup size after the windows is resized to fullscreen. Cheers, Anders I reckon you are on the right track, Anders. I have already narrowed narrowed the problem down to fgdata, as building next of yesterday against with an older fgdata checkout works just fine. I shall fast forward to the commit you suggest. Grr, I am beginning to enjoy this here git, it's cool - even for a senile old bugger like me :-) Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
Can't think how to proceed from here though :-( Sigh! As you and others have mentioned, the wxradar connection ought to be a clue, but I am not brainy enough anymore to follow it up, especially if involves delving into OSG. If you are able to start Rembrandt, does it change anything ? Regards, -Fred -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On Tue, 2012-09-04 at 18:00 +0200, Frederic Bouvier wrote: Can't think how to proceed from here though :-( Sigh! As you and others have mentioned, the wxradar connection ought to be a clue, but I am not brainy enough anymore to follow it up, especially if involves delving into OSG. If you are able to start Rembrandt, does it change anything ? Regards, -Fred You bet your cotton-picking socks it does! Bravo! Horrible display, but that's to be expected on naff built-in graphics, without tuning: https://dl.dropbox.com/u/15936159/fgfs-rembrandt.png Shows full screen + screwy view port Progress ?? Alasdair :-) -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On Tue, 2012-09-04 at 18:00 +0200, Frederic Bouvier wrote: Can't think how to proceed from here though :-( Sigh! As you and others have mentioned, the wxradar connection ought to be a clue, but I am not brainy enough anymore to follow it up, especially if involves delving into OSG. If you are able to start Rembrandt, does it change anything ? Regards, -Fred You bet your cotton-picking socks it does! Bravo! Horrible display, but that's to be expected on naff built-in graphics, without tuning: https://dl.dropbox.com/u/15936159/fgfs-rembrandt.png Shows full screen + screwy view port Progress ?? And the same screenshot with 'Debug Hide/Show Rendering Buffers' menu on, like in http://wiki.flightgear.org/File:Project_rembrandt_1.png (notice the corners) ? Regards, -Fred -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On Tue, 2012-09-04 at 18:52 +0200, Frederic Bouvier wrote: On Tue, 2012-09-04 at 18:00 +0200, Frederic Bouvier wrote: Can't think how to proceed from here though :-( Sigh! As you and others have mentioned, the wxradar connection ought to be a clue, but I am not brainy enough anymore to follow it up, especially if involves delving into OSG. If you are able to start Rembrandt, does it change anything ? Regards, -Fred You bet your cotton-picking socks it does! Bravo! Horrible display, but that's to be expected on naff built-in graphics, without tuning: https://dl.dropbox.com/u/15936159/fgfs-rembrandt.png Shows full screen + screwy view port Progress ?? And the same screenshot with 'Debug Hide/Show Rendering Buffers' menu on, like in http://wiki.flightgear.org/File:Project_rembrandt_1.png (notice the corners) ? Regards, -Fred Here you go: https://dl.dropbox.com/u/15936159/fgfs-rem_rend_buffers_on.png You seem to be scenting the fox here, Fred Wish I knew what the rendering buffers tell you! Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
Here you go: https://dl.dropbox.com/u/15936159/fgfs-rem_rend_buffers_on.png You seem to be scenting the fox here, Fred Wish I knew what the rendering buffers tell you! The corners are rendered in a first pass in textures. These textures are resized when an event from OSG tells to, and then a transformation tells OpenGL to take the whole allocated space. Here, we see that the corners are not covering their allocated space, certainly because a bad viewport is given. The green image in the main window is in fact a Rembrandt bug (and a known source of bad performance) that is hidden when the images in the corners occupy all there space. It has the right viewport though. But without being able to reproduce, I don't know the reason of that. Regards, -Fred -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
Here you go: https://dl.dropbox.com/u/15936159/fgfs-rem_rend_buffers_on.png You seem to be scenting the fox here, Fred Wish I knew what the rendering buffers tell you! The corners are rendered in a first pass in textures. These textures are resized when an event from OSG tells to, and then a transformation tells OpenGL to take the whole allocated space. Here, we see that the corners are not covering their allocated space, certainly because a bad viewport is given. The green image in the main window is in fact a Rembrandt bug (and a known source of bad performance) that is hidden when the images in the corners occupy all there space. It has the right viewport though. But without being able to reproduce, I don't know the reason of that. BTW: it looks like this one : http://code.google.com/p/flightgear-bugs/issues/detail?id=354 -Fred -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
Here you go: https://dl.dropbox.com/u/15936159/fgfs-rem_rend_buffers_on.png You seem to be scenting the fox here, Fred Wish I knew what the rendering buffers tell you! The corners are rendered in a first pass in textures. These textures are resized when an event from OSG tells to, and then a transformation tells OpenGL to take the whole allocated space. Here, we see that the corners are not covering their allocated space, certainly because a bad viewport is given. The green image in the main window is in fact a Rembrandt bug (and a known source of bad performance) that is hidden when the images in the corners occupy all there space. It has the right viewport though. But without being able to reproduce, I don't know the reason of that. BTW: it looks like this one : http://code.google.com/p/flightgear-bugs/issues/detail?id=354 If you start FG with --log-level=debug --log-class=view you'll be able to see messages like : FGEventHandler::handle: RESIZE event 763 x 1338, resizable: 1 FGRenderer::resize: new size 1338 x 763 What are yours ? Regards, -Fred -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
If you start FG with --log-level=debug --log-class=view you'll be able to see messages like : FGEventHandler::handle: RESIZE event 763 x 1338, resizable: 1 FGRenderer::resize: new size 1338 x 763 What are yours ? Regards, -Fred I see FGRenderer::resize: new size 1904 x 993 but no reference to FGEventHandler::? I already has a look at FGRenderer::resize, but the method doesn't seem to do anything, except issue the above message at debug level. Keep going, Fred! I have every confidence you can crack this problem. Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
If you start FG with --log-level=debug --log-class=view you'll be able to see messages like : FGEventHandler::handle: RESIZE event 763 x 1338, resizable: 1 FGRenderer::resize: new size 1338 x 763 What are yours ? Regards, -Fred I see FGRenderer::resize: new size 1904 x 993 but no reference to FGEventHandler::? I already has a look at FGRenderer::resize, but the method doesn't seem to do anything, except issue the above message at debug level. Keep going, Fred! I have every confidence you can crack this problem. It's the FGEventHandler::handle that actually do the resize. So if you don't see it it explains what you see. BTW: What is your OSG version ? Regards, -Fred -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On Tue, 2012-09-04 at 22:34 +0200, Frederic Bouvier wrote: If you start FG with --log-level=debug --log-class=view you'll be able to see messages like : FGEventHandler::handle: RESIZE event 763 x 1338, resizable: 1 FGRenderer::resize: new size 1338 x 763 What are yours ? Regards, -Fred I see FGRenderer::resize: new size 1904 x 993 but no reference to FGEventHandler::? I already has a look at FGRenderer::resize, but the method doesn't seem to do anything, except issue the above message at debug level. Keep going, Fred! I have every confidence you can crack this problem. It's the FGEventHandler::handle that actually do the resize. So if you don't see it it explains what you see. BTW: What is your OSG version ? Regards, -Fred 3.0.1 I always used my own build, but also tried debian's deb (version 3.0.1-3 +b2) Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On Tue, 2012-09-04 at 21:58 +0100, Alasdair wrote: On Tue, 2012-09-04 at 22:34 +0200, Frederic Bouvier wrote: If you start FG with --log-level=debug --log-class=view you'll be able to see messages like : FGEventHandler::handle: RESIZE event 763 x 1338, resizable: 1 FGRenderer::resize: new size 1338 x 763 What are yours ? Regards, -Fred I see FGRenderer::resize: new size 1904 x 993 but no reference to FGEventHandler::? I already has a look at FGRenderer::resize, but the method doesn't seem to do anything, except issue the above message at debug level. Keep going, Fred! I have every confidence you can crack this problem. It's the FGEventHandler::handle that actually do the resize. So if you don't see it it explains what you see. BTW: What is your OSG version ? Regards, -Fred 3.0.1 I always used my own build, but also tried debian's deb (version 3.0.1-3 +b2) Alasdair I see : FGEventHandler::handle: RESIZE event 1002 x 1920, resizable: 1 FGRenderer::resize: new size 1920 x 1002 but _only after I click the maximize widget_ Screen reverts to small view if I click any menu item Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On Mon, 2012-09-03 at 01:06 +0200, ThorstenB wrote: On Sun, Sep 2, 2012 at 11:31 AM, Alasdair ali...@btinternet.com wrote: After git pull (sg,fg,fgdata) a couple of days ago, I was presented with a strange screen which I had not seen before. My regular was shown as a sub-screen in the bottom-left corner. Clicking the maximise widget restores full screen, but any click on a menu item returns me to this stupid screen. Is this infuriating behaviour deliberate, or has my graphics driver gone walk-about? https://dl.dropbox.com/u/15936159/fg_quarter-screen.png Unfortunately a long standing issue. This affects ATI graphics cards when running newer drivers. It can be fixed by reverting to an older ATI driver (see link below). To me it looked like a genuine OSG or ATI driver bug (the size of the graphics view port doesn't properly resize to the size of the window), but it's likely triggered by something special we're doing in FG, since other applications don't seem to have this issue. Also, it's timing-/sequence dependant: it only affects certain aircraft (aircraft using wxradar, which changes/delays the OSG init sequence on startup, are less likely to have the issue) - and for some people it dis- and reappears with certain FG builds. It would be really good if someone seeing the issue locally could properly debug this - even if it means digging into OSG - since a larger number of people are affected. I tried remote debugging this a while ago, but gave up after being unable to narrow it down inside FG. Maybe someone else has better luck. For more see here: http://code.google.com/p/flightgear-bugs/issues/detail?id=385 cheers, Thorsten Thanks very much for the link, Thorsten, it is a good start. But strange that this problem, should just appear without any driver change. But I am using Debian SID, and update my system daily, so it could be OS-related. I will install UBUNTU and try the same build under that OS. regards, Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
Im running Ubuntu , still get that problem.I've added the radar to my intrumentation file in my affected aircraft , and the problem is gone.Strange. Syd On Mon, Sep 3, 2012 at 2:23 AM, Alasdair ali...@btinternet.com wrote: On Mon, 2012-09-03 at 01:06 +0200, ThorstenB wrote: On Sun, Sep 2, 2012 at 11:31 AM, Alasdair ali...@btinternet.com wrote: After git pull (sg,fg,fgdata) a couple of days ago, I was presented with a strange screen which I had not seen before. My regular was shown as a sub-screen in the bottom-left corner. Clicking the maximise widget restores full screen, but any click on a menu item returns me to this stupid screen. Is this infuriating behaviour deliberate, or has my graphics driver gone walk-about? https://dl.dropbox.com/u/15936159/fg_quarter-screen.png Unfortunately a long standing issue. This affects ATI graphics cards when running newer drivers. It can be fixed by reverting to an older ATI driver (see link below). To me it looked like a genuine OSG or ATI driver bug (the size of the graphics view port doesn't properly resize to the size of the window), but it's likely triggered by something special we're doing in FG, since other applications don't seem to have this issue. Also, it's timing-/sequence dependant: it only affects certain aircraft (aircraft using wxradar, which changes/delays the OSG init sequence on startup, are less likely to have the issue) - and for some people it dis- and reappears with certain FG builds. It would be really good if someone seeing the issue locally could properly debug this - even if it means digging into OSG - since a larger number of people are affected. I tried remote debugging this a while ago, but gave up after being unable to narrow it down inside FG. Maybe someone else has better luck. For more see here: http://code.google.com/p/flightgear-bugs/issues/detail?id=385 cheers, Thorsten Thanks very much for the link, Thorsten, it is a good start. But strange that this problem, should just appear without any driver change. But I am using Debian SID, and update my system daily, so it could be OS-related. I will install UBUNTU and try the same build under that OS. regards, Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
Hi Alasdair, I also do not think it is OS related, since I have seen this 'sometimes' in Windows 7 64-bit, with AMD Radeon HD 7700 Series with (I hope) the latest Catalyst driver! At the time I thought little of it since it seemed to go away a few seconds later... by itself... In the Catalyst Control Center it shows a Driver Package Version 8.961-120405a-137813C-ATI, Catalyst Version 12.1, dated 29-Feb-2012... Hmmm, contrary to last time I checked for updates, it does now report a later Catalyst version 12.8 available... will download and try that... Now shows driver 8.982-120727a-145524C-ATI in Catalyst 12.8, dated today... Just tried FG 2.8.9.4 with 777-200 and no problem, but as indicated it only happened 'sometimes' and was never able to isolate the specific scenario that caused it... maybe as you indicate it is a/c specific... Found any reason why adding the radar to the aircraft made it go away... Will keep testing with different aircraft... Regards, Geoff. On Mon, 2012-09-03 at 09:39 -0600, syd adams wrote: Im running Ubuntu , still get that problem.I've added the radar to my intrumentation file in my affected aircraft , and the problem is gone.Strange. Syd On Mon, Sep 3, 2012 at 2:23 AM, Alasdair ali...@btinternet.com wrote: On Mon, 2012-09-03 at 01:06 +0200, ThorstenB wrote: On Sun, Sep 2, 2012 at 11:31 AM, Alasdair ali...@btinternet.com wrote: After git pull (sg,fg,fgdata) a couple of days ago, I was presented with a strange screen which I had not seen before. My regular was shown as a sub-screen in the bottom-left corner. Clicking the maximise widget restores full screen, but any click on a menu item returns me to this stupid screen. Is this infuriating behaviour deliberate, or has my graphics driver gone walk-about? https://dl.dropbox.com/u/15936159/fg_quarter-screen.png Unfortunately a long standing issue. This affects ATI graphics cards when running newer drivers. It can be fixed by reverting to an older ATI driver (see link below). To me it looked like a genuine OSG or ATI driver bug (the size of the graphics view port doesn't properly resize to the size of the window), but it's likely triggered by something special we're doing in FG, since other applications don't seem to have this issue. Also, it's timing-/sequence dependant: it only affects certain aircraft (aircraft using wxradar, which changes/delays the OSG init sequence on startup, are less likely to have the issue) - and for some people it dis- and reappears with certain FG builds. It would be really good if someone seeing the issue locally could properly debug this - even if it means digging into OSG - since a larger number of people are affected. I tried remote debugging this a while ago, but gave up after being unable to narrow it down inside FG. Maybe someone else has better luck. For more see here: http://code.google.com/p/flightgear-bugs/issues/detail?id=385 cheers, Thorsten Thanks very much for the link, Thorsten, it is a good start. But strange that this problem, should just appear without any driver change. But I am using Debian SID, and update my system daily, so it could be OS-related. I will install UBUNTU and try the same build under that OS. regards, Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On Mon, 2012-09-03 at 09:39 -0600, syd adams wrote: Im running Ubuntu , still get that problem.I've added the radar to my intrumentation file in my affected aircraft , and the problem is gone.Strange. Syd On Mon, Sep 3, 2012 at 2:23 AM, Alasdair ali...@btinternet.com wrote: On Mon, 2012-09-03 at 01:06 +0200, ThorstenB wrote: On Sun, Sep 2, 2012 at 11:31 AM, Alasdair ali...@btinternet.com wrote: After git pull (sg,fg,fgdata) a couple of days ago, I was presented with a strange screen which I had not seen before. My regular was shown as a sub-screen in the bottom-left corner. Clicking the maximise widget restores full screen, but any click on a menu item returns me to this stupid screen. Is this infuriating behaviour deliberate, or has my graphics driver gone walk-about? https://dl.dropbox.com/u/15936159/fg_quarter-screen.png Unfortunately a long standing issue. This affects ATI graphics cards when running newer drivers. It can be fixed by reverting to an older ATI driver (see link below). To me it looked like a genuine OSG or ATI driver bug (the size of the graphics view port doesn't properly resize to the size of the window), but it's likely triggered by something special we're doing in FG, since other applications don't seem to have this issue. Also, it's timing-/sequence dependant: it only affects certain aircraft (aircraft using wxradar, which changes/delays the OSG init sequence on startup, are less likely to have the issue) - and for some people it dis- and reappears with certain FG builds. It would be really good if someone seeing the issue locally could properly debug this - even if it means digging into OSG - since a larger number of people are affected. I tried remote debugging this a while ago, but gave up after being unable to narrow it down inside FG. Maybe someone else has better luck. For more see here: http://code.google.com/p/flightgear-bugs/issues/detail?id=385 cheers, Thorsten Thanks very much for the link, Thorsten, it is a good start. But strange that this problem, should just appear without any driver change. But I am using Debian SID, and update my system daily, so it could be OS-related. I will install UBUNTU and try the same build under that OS. regards, Alasdair OK I tried building the same git version on Ubuntu (different gcc version) Same problem. Tried a complete fresh clone (sg/fg/fgdata) Same problem. Downloaded and built 2.8.0 and built with same script : NO problem. So this is a git problem and fairly recent too. 0- Most aircraft I have tested behave properly, but the ones that have this strange viewport problem include : c172p, b1900d, SenecaII. The viewport changes immediately after the splash screen displays initializing graphics engine. Maybe someone can figure out what these have in common that the 707, 737-300, 777-200ER, A-10, A300, aerostar700, alphajet, b29, bf109g, Bravo, c310, Citation*, CRJ700ER, Cub and pa24-250-CIIB do not. I cannot therefore attribute this problem to the Catalyst driver. My git pull/build of 12 Aug did not display this problem. That of 1 Sep did. So I reckon it must have been a commit made between these dates. Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
I cannot therefore attribute this problem to the Catalyst driver. My git pull/build of 12 Aug did not display this problem. That of 1 Sep did. So I reckon it must have been a commit made between these dates. So you may want to git bisect to find which commit was the problem Thanks, -Fred -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On Mon, 2012-09-03 at 20:22 +0200, Geoff McLane wrote: Hi Alasdair, I also do not think it is OS related, since I have seen this 'sometimes' in Windows 7 64-bit, with AMD Radeon HD 7700 Series with (I hope) the latest Catalyst driver! At the time I thought little of it since it seemed to go away a few seconds later... by itself... In the Catalyst Control Center it shows a Driver Package Version 8.961-120405a-137813C-ATI, Catalyst Version 12.1, dated 29-Feb-2012... Hmmm, contrary to last time I checked for updates, it does now report a later Catalyst version 12.8 available... will download and try that... Now shows driver 8.982-120727a-145524C-ATI in Catalyst 12.8, dated today... Just tried FG 2.8.9.4 with 777-200 and no problem, but as indicated it only happened 'sometimes' and was never able to isolate the specific scenario that caused it... maybe as you indicate it is a/c specific... Found any reason why adding the radar to the aircraft made it go away... Will keep testing with different aircraft... Regards, Geoff. Geoff, thanks for the input. My catalyst version is 12.6. But the 777-200ER did not exhibit this behaviour for me.. just a limited set planes, see my post of a few minutes ago. Alasdair On Mon, 2012-09-03 at 09:39 -0600, syd adams wrote: Im running Ubuntu , still get that problem.I've added the radar to my intrumentation file in my affected aircraft , and the problem is gone.Strange. Syd On Mon, Sep 3, 2012 at 2:23 AM, Alasdair ali...@btinternet.com wrote: On Mon, 2012-09-03 at 01:06 +0200, ThorstenB wrote: On Sun, Sep 2, 2012 at 11:31 AM, Alasdair ali...@btinternet.com wrote: After git pull (sg,fg,fgdata) a couple of days ago, I was presented with a strange screen which I had not seen before. My regular was shown as a sub-screen in the bottom-left corner. Clicking the maximise widget restores full screen, but any click on a menu item returns me to this stupid screen. Is this infuriating behaviour deliberate, or has my graphics driver gone walk-about? https://dl.dropbox.com/u/15936159/fg_quarter-screen.png Unfortunately a long standing issue. This affects ATI graphics cards when running newer drivers. It can be fixed by reverting to an older ATI driver (see link below). To me it looked like a genuine OSG or ATI driver bug (the size of the graphics view port doesn't properly resize to the size of the window), but it's likely triggered by something special we're doing in FG, since other applications don't seem to have this issue. Also, it's timing-/sequence dependant: it only affects certain aircraft (aircraft using wxradar, which changes/delays the OSG init sequence on startup, are less likely to have the issue) - and for some people it dis- and reappears with certain FG builds. It would be really good if someone seeing the issue locally could properly debug this - even if it means digging into OSG - since a larger number of people are affected. I tried remote debugging this a while ago, but gave up after being unable to narrow it down inside FG. Maybe someone else has better luck. For more see here: http://code.google.com/p/flightgear-bugs/issues/detail?id=385 cheers, Thorsten Thanks very much for the link, Thorsten, it is a good start. But strange that this problem, should just appear without any driver change. But I am using Debian SID, and update my system daily, so it could be OS-related. I will install UBUNTU and try the same build under that OS. regards, Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On Mon, 2012-09-03 at 22:40 +0200, Frederic Bouvier wrote: I cannot therefore attribute this problem to the Catalyst driver. My git pull/build of 12 Aug did not display this problem. That of 1 Sep did. So I reckon it must have been a commit made between these dates. So you may want to git bisect to find which commit was the problem Thanks, -Fred Where can I learn how to do that, Fred? My knowledge of git ends with clone/pull. Suggested reading? regards, Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On Mon, 03 Sep 2012 21:32:02 +0100, Alasdair wrote in message 1346704322.26230.28.camel@executrix: I cannot therefore attribute this problem to the Catalyst driver. ..one quick way to find out, is fetch a live usb distro, e.g. Debian Live, (there's also Ubuntu, Knoppix etc) put it on an usb key and boot it, it should run the free open source radeon driver, not ATI's fglrx. ..then do a build from your git tree with either Geoff or Brisa's build script, run it and see what happens. ..a third way is run your built binaries thru Wine. ..I'm looking for a way to boot iso etc images from xen and I believe I'm missing something stupid simple, hints or ideas, please? My git pull/build of 12 Aug did not display this problem. That of 1 Sep did. So I reckon it must have been a commit made between these dates. -- ..med vennlig hilsen = with Kind Regards from Arnt Karlsen ...with a number of polar bear hunters in his ancestry... Scenarios always come in sets of three: best case, worst case, and just in case. -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
Hi Alasdair, This tutorial on git bisect has references to Drupal but is otherwise very clear: http://webchick.net/node/99 You basically start from a known-good point, a known-bad one and then tell git bisect if the next checkouts are good (no problem) or bad (show the problem). Each test requires recompilation of FG and possibly multiple runs if the problem is intermittent. Hope this helps, Tom On Mon, Sep 3, 2012 at 1:56 PM, Alasdair ali...@btinternet.com wrote: On Mon, 2012-09-03 at 22:40 +0200, Frederic Bouvier wrote: I cannot therefore attribute this problem to the Catalyst driver. My git pull/build of 12 Aug did not display this problem. That of 1 Sep did. So I reckon it must have been a commit made between these dates. So you may want to git bisect to find which commit was the problem Thanks, -Fred Where can I learn how to do that, Fred? My knowledge of git ends with clone/pull. Suggested reading? regards, Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
Ive had this same issue for months now , on a laptop with ATI Mobility Radeon HD 4250.It only happens on some aircraft , but i havent really tried to hunt the problem down. Im suspecting a shortage of video memory . I also can get full screen by changing the window size , but any dialog or text on the screen reduces it back to the bottom left corner.Possibly related , Ive noticed that the aspect ratio isnt updated when resizing the window. I hope someone else has an idea where to start debugging this . Syd On Sun, Sep 2, 2012 at 11:31 AM, Alasdair ali...@btinternet.com wrote: After git pull (sg,fg,fgdata) a couple of days ago, I was presented with a strange screen which I had not seen before. My regular was shown as a sub-screen in the bottom-left corner. Clicking the maximise widget restores full screen, but any click on a menu item returns me to this stupid screen. Is this infuriating behaviour deliberate, or has my graphics driver gone walk-about? https://dl.dropbox.com/u/15936159/fg_quarter-screen.png Kind regards, Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On Sun, 2012-09-02 at 12:26 -0600, syd adams wrote: Ive had this same issue for months now , on a laptop with ATI Mobility Radeon HD 4250.It only happens on some aircraft , but i havent really tried to hunt the problem down. Im suspecting a shortage of video memory . I also can get full screen by changing the window size , but any dialog or text on the screen reduces it back to the bottom left corner.Possibly related , Ive noticed that the aspect ratio isnt updated when resizing the window. I hope someone else has an idea where to start debugging this . Syd On Sun, Sep 2, 2012 at 11:31 AM, Alasdair ali...@btinternet.com wrote: After git pull (sg,fg,fgdata) a couple of days ago, I was presented with a strange screen which I had not seen before. My regular was shown as a sub-screen in the bottom-left corner. Clicking the maximise widget restores full screen, but any click on a menu item returns me to this stupid screen. Is this infuriating behaviour deliberate, or has my graphics driver gone walk-about? https://dl.dropbox.com/u/15936159/fg_quarter-screen.png Kind regards, Alasdair This is also an a Radeon device (HD 6530D) on my quiet machine, no moving parts except for cpu fan. But this is certainly a NEW problem, cos I have had no problems since yesterdays git build, last git pull about a week ago. I will check it out on my big (noisy) machine with Nvidia cards, but not till I can backup my working fg. Thanks for your interest, Syd Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
Please open a bug for this, if there is not one already (it sounds similar to one existing issue), and include all the information below - especially when it broke, OSG versions, graphics cards and so on. Cheers, James On 2 Sep 2012, at 20:07, Alasdair ali...@btinternet.com wrote: On Sun, 2012-09-02 at 12:26 -0600, syd adams wrote: Ive had this same issue for months now , on a laptop with ATI Mobility Radeon HD 4250.It only happens on some aircraft , but i havent really tried to hunt the problem down. Im suspecting a shortage of video memory . I also can get full screen by changing the window size , but any dialog or text on the screen reduces it back to the bottom left corner.Possibly related , Ive noticed that the aspect ratio isnt updated when resizing the window. I hope someone else has an idea where to start debugging this . Syd On Sun, Sep 2, 2012 at 11:31 AM, Alasdair ali...@btinternet.com wrote: After git pull (sg,fg,fgdata) a couple of days ago, I was presented with a strange screen which I had not seen before. My regular was shown as a sub-screen in the bottom-left corner. Clicking the maximise widget restores full screen, but any click on a menu item returns me to this stupid screen. Is this infuriating behaviour deliberate, or has my graphics driver gone walk-about? https://dl.dropbox.com/u/15936159/fg_quarter-screen.png Kind regards, Alasdair This is also an a Radeon device (HD 6530D) on my quiet machine, no moving parts except for cpu fan. But this is certainly a NEW problem, cos I have had no problems since yesterdays git build, last git pull about a week ago. I will check it out on my big (noisy) machine with Nvidia cards, but not till I can backup my working fg. Thanks for your interest, Syd Alasdair -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] strange screen
On Sun, Sep 2, 2012 at 11:31 AM, Alasdair ali...@btinternet.com wrote: After git pull (sg,fg,fgdata) a couple of days ago, I was presented with a strange screen which I had not seen before. My regular was shown as a sub-screen in the bottom-left corner. Clicking the maximise widget restores full screen, but any click on a menu item returns me to this stupid screen. Is this infuriating behaviour deliberate, or has my graphics driver gone walk-about? https://dl.dropbox.com/u/15936159/fg_quarter-screen.png Unfortunately a long standing issue. This affects ATI graphics cards when running newer drivers. It can be fixed by reverting to an older ATI driver (see link below). To me it looked like a genuine OSG or ATI driver bug (the size of the graphics view port doesn't properly resize to the size of the window), but it's likely triggered by something special we're doing in FG, since other applications don't seem to have this issue. Also, it's timing-/sequence dependant: it only affects certain aircraft (aircraft using wxradar, which changes/delays the OSG init sequence on startup, are less likely to have the issue) - and for some people it dis- and reappears with certain FG builds. It would be really good if someone seeing the issue locally could properly debug this - even if it means digging into OSG - since a larger number of people are affected. I tried remote debugging this a while ago, but gave up after being unable to narrow it down inside FG. Maybe someone else has better luck. For more see here: http://code.google.com/p/flightgear-bugs/issues/detail?id=385 cheers, Thorsten -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
