Re: [Flightgear-devel] pilot list...
> After Torsten's change to multiplayer.nas , Im seeing similar errors ... > just haven't had a chance to get back too looking at it... After looking at the code again, I can't see a link between the two issues. But I also don't feel to much at home in the AI code, so I probably (likely) overlooked something. Torsten -- Come build with us! The BlackBerry® Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9-12, 2009. Register now! http://p.sf.net/sfu/devconf ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] pilot list...
After Torsten's change to multiplayer.nas , Im seeing similar errors ... just haven't had a chance to get back too looking at it... -- Come build with us! The BlackBerry® Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9-12, 2009. Register now! http://p.sf.net/sfu/devconf___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] pilot list...
On Thu, 17 Sep 2009, Nicolas Quijano wrote: > Still a problem, with or without the MP ignore chat. (tested under both > conditions) > We intermittently now get chat messages from people thousands of kilometres > away and duplicated messages with the callsign changed : for example, > some_nic says : nice barrel roll, then some_other_nic a few seconds later > will "magically" repeat what was just said, absolutely identical, > punctuation and typos included :) The messages from half a world away can AFAICS only be due to a breakdown of the distance check in fgms - otherwise they'd never reach your FlightGear instance. However, such a breakage could be due to a bug in fgms (e.g. in the decoding of the position) or due to the sending FlightGear instance sending bogus data. I have also seen these far away messages but have assumed that I got them due to the sender having NaN in his position (i.e. the sender's FG crashed but managed to fire off a last packet with messed up position data). I have no real evidence that that really is the case, though. Running with FPE traps enabled or instrumenting the code in multiplayer.nas might give some more information. > Log on messages are also a bit on the fritz : you can see this happens with > individuals log-on messages being spoken by other nics : afaik, only someguy > uses Phhtt but I saw among others, syd sign on with that message ;) Log on messages are just like any MP chat message. I recall having seen occasional misattributed MP chat messages for a long time, so I don't think that is a new problem. (Typically I'd react to mpdummy saying things (though in some cases that could be a callsign clash). As for the problem of the late initialization of the valid property (compared to the model-added property): How about the listener just setting an update flag which cause the pilot dialog to regenerate the pilot list the next time it runs? That way the arrival of a new user will appear atomic to the dialog update code (since it is only the listener that is run from inside the C++ setValue() call). Cheers, Anders -- --- Anders Gidenstam WWW: http://www.gidenstam.org/FlightGear/ -- Come build with us! The BlackBerry® Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9-12, 2009. Register now! http://p.sf.net/sfu/devconf ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] pilot list...
Still a problem, with or without the MP ignore chat. (tested under both conditions) We intermittently now get chat messages from people thousands of kilometres away and duplicated messages with the callsign changed : for example, some_nic says : nice barrel roll, then some_other_nic a few seconds later will "magically" repeat what was just said, absolutely identical, punctuation and typos included :) It kinda makes text based ATC a nightmare, as it looks random to the user busy with flying. Log on messages are also a bit on the fritz : you can see this happens with individuals log-on messages being spoken by other nics : afaik, only someguy uses Phhtt but I saw among others, syd sign on with that message ;) Or just a few minutes ago, while I was getting ready to launch from Nimitz, a guy in Europe in a F-18 was asking for partners to fly to the Balkans, and I thought he was nuts to want to fly from KSFO to Montenegro, and couldn't see him on my zoomed mpmap. Of course not : zoomed out and he was in Europe, near his intended destination. Common occurrence in the last couple weeks. Probably the "last in line" bug affecting the pilot list and chat in a weird way. Cheers, Nic On Sun, Aug 30, 2009 at 5:19 PM, Rob Shearman, Jr. wrote: > I've mentioned the problem here before, but as far as I know, no one has > pinned it down, because all of the Win32 builds from June forward exhibit > this behavior. Thanks for looking into it! -R. > > Robert M. Shearman, Jr. > Transit Operations Supervisor, > University of Maryland Department of Transportation > also known as rm...@umd.edu > > -- > *From:* syd adams > *To:* flightgear-dev > *Sent:* Sunday, August 30, 2009 1:24:13 AM > *Subject:* [Flightgear-devel] pilot list... > > While attempting to hunt down a chat message problem , (no messages > displayed from certain players), > I found that the pilot list is always 1 player short , and the "model view" > skips that player. > They all appear in ai/models , but num-players is always show 1 less > player. > I've been sifting through the multiplay.nas , but thought I'd mention it > here in case someone has > already looked at the problem . > cheers > > > > > -- > Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day > trial. Simplify your report design, integration and deployment - and focus > on > what you do best, core application coding. Discover what's new with > Crystal Reports now. http://p.sf.net/sfu/bobj-july > ___ > Flightgear-devel mailing list > Flightgear-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/flightgear-devel > > -- Be Kind. Remember, everyone is fighting a hard battle. -- Come build with us! The BlackBerry® Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9-12, 2009. Register now! http://p.sf.net/sfu/devconf___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] pilot list error
Torsten Dreyer wrote: > Looks like this is a chicken-egg problem: > > The chicken: > The model class in multiplayer.nas listens on the > property /ai/models/model-added and checks for all models the > property "valid" to be true. > > The egg: > When AIManager attaches a model, it first loads the model by calling > model->init(); > FGAIBase::init() calls FGAIBase::load3DModel() which calls > FGAIBase::initModel() which sets property /ai/models/model-added. (This > triggers the chicken, but "valid" is not yet set to true) > _AFTER_ model->init() in AIManager::atach() is finished, the property "valid" > in the /ai/models/multiplayer[n] is set to true. > > This causes new latest addition to the multiplayer list to be ignored. > > I think the behaviour of FGAIBase and AIManager is correct to set the "valid" > property to true after the initialization is complete. > I also think that the check for "valid" in the model class in multiplayer.nas > is required to stay there. > > I have just commited a patch for multiplayer.nas that ignores the "valid" > flag > for the latest pilot that has joined. > > Please check if this solves the problem. > > Torsten Thanks for working this out Torsten. I suspected that the problem was related to the valid property, but didn't manage to get to the bottom of it. Is there any mileage in only setting the model-added property in the AIManager::attach() function after mode_init() is complete? -Stuart (been away for two weeks so only just catching up on developments) -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] pilot list error
Last test shows correct amount of players , pilot list good , model view good , but now apparently when someone leaves and valid becomes false , they remain on the pilot list / model view for a short time , though no 3d model is visible in the model view. But I need sleep , have to look at it tomorrow. Cheers On Sat, Sep 5, 2009 at 10:44 PM, syd adams wrote: > Changing line 275 in multiplayer.nas to > > "distance-to-km": distance / 1000.0 or 0, > > appears to get things running again. > > > -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] pilot list error
Changing line 275 in multiplayer.nas to "distance-to-km": distance / 1000.0 or 0, appears to get things running again. -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] pilot list error
After the first test I'm getting ... Nasal runtime error: nil used in numeric context at /home/syd/FGFS/data/Nasal/multiplayer.nas, line 275 which of coarse causes the pilot list to not appear. Will continue testing . -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] pilot list error
Thanks for looking into it.Torsten , just got home from a soggy camping trip , so I can dry out and test your changes :) cheers On Sat, Sep 5, 2009 at 1:22 PM, Torsten Dreyer wrote: > > As I mentioned before , I've been looking for the missing player in the > > pilot list and model view. > > If I comment out the 2 lines below "foreach" , in multiplayer.nas at > line > > 374 ... > > > > foreach (var n; props.globals.getNode("ai/models", > > 1).getChildren("multiplayer")) { > > #if (!n.getNode("valid", 1).getValue()) > > # continue; > > > > ... I then get the correct in /ai/models, and I can now > step > > through ALL the players in "model view" , > > and they all appear in the pilot list. > > I don't consider this a fix since I dont know why the above lines would > > cause a problem,but it's a start. > > Cheers > Looks like this is a chicken-egg problem: > > The chicken: > The model class in multiplayer.nas listens on the > property /ai/models/model-added and checks for all models the > property "valid" to be true. > > The egg: > When AIManager attaches a model, it first loads the model by calling > model->init(); > FGAIBase::init() calls FGAIBase::load3DModel() which calls > FGAIBase::initModel() which sets property /ai/models/model-added. (This > triggers the chicken, but "valid" is not yet set to true) > _AFTER_ model->init() in AIManager::atach() is finished, the property > "valid" > in the /ai/models/multiplayer[n] is set to true. > > This causes new latest addition to the multiplayer list to be ignored. > > I think the behaviour of FGAIBase and AIManager is correct to set the > "valid" > property to true after the initialization is complete. > I also think that the check for "valid" in the model class in > multiplayer.nas > is required to stay there. > > I have just commited a patch for multiplayer.nas that ignores the "valid" > flag > for the latest pilot that has joined. > > Please check if this solves the problem. > > Torsten > > > -- > Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day > trial. Simplify your report design, integration and deployment - and focus > on > what you do best, core application coding. Discover what's new with > Crystal Reports now. http://p.sf.net/sfu/bobj-july > ___ > Flightgear-devel mailing list > Flightgear-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/flightgear-devel > -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] pilot list error
On 09/05/09 13:22, Torsten Dreyer wrote: > Looks like this is a chicken-egg problem: > > The chicken: > The model class in multiplayer.nas listens on the > property /ai/models/model-added and checks for all models the > property "valid" to be true. > > The egg: > When AIManager attaches a model, it first loads the model by calling > model->init(); > FGAIBase::init() calls FGAIBase::load3DModel() which calls > FGAIBase::initModel() which sets property /ai/models/model-added. (This > triggers the chicken, but "valid" is not yet set to true) > _AFTER_ model->init() in AIManager::atach() is finished, the property "valid" > in the /ai/models/multiplayer[n] is set to true. > > This causes new latest addition to the multiplayer list to be ignored. > > I think the behaviour of FGAIBase and AIManager is correct to set the "valid" > property to true after the initialization is complete. > I also think that the check for "valid" in the model class in multiplayer.nas > is required to stay there. > > I have just commited a patch for multiplayer.nas that ignores the "valid" > flag > for the latest pilot that has joined. > > Please check if this solves the problem. The conventional technique for dealing with such problems is to have three states: 00 invalid 01 incipient, i.e. usable for some purposes 10 fully_valid, i.e. usable for all purposes Initialization routines typically require incipient or fully_valid, while other routines typically insist on fully_valid. This is less risky and more maintainable than simply ignoring the valid flag. -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] pilot list error
> As I mentioned before , I've been looking for the missing player in the > pilot list and model view. > If I comment out the 2 lines below "foreach" , in multiplayer.nas at line > 374 ... > > foreach (var n; props.globals.getNode("ai/models", > 1).getChildren("multiplayer")) { > #if (!n.getNode("valid", 1).getValue()) > # continue; > > ... I then get the correct in /ai/models, and I can now step > through ALL the players in "model view" , > and they all appear in the pilot list. > I don't consider this a fix since I dont know why the above lines would > cause a problem,but it's a start. > Cheers Looks like this is a chicken-egg problem: The chicken: The model class in multiplayer.nas listens on the property /ai/models/model-added and checks for all models the property "valid" to be true. The egg: When AIManager attaches a model, it first loads the model by calling model->init(); FGAIBase::init() calls FGAIBase::load3DModel() which calls FGAIBase::initModel() which sets property /ai/models/model-added. (This triggers the chicken, but "valid" is not yet set to true) _AFTER_ model->init() in AIManager::atach() is finished, the property "valid" in the /ai/models/multiplayer[n] is set to true. This causes new latest addition to the multiplayer list to be ignored. I think the behaviour of FGAIBase and AIManager is correct to set the "valid" property to true after the initialization is complete. I also think that the check for "valid" in the model class in multiplayer.nas is required to stay there. I have just commited a patch for multiplayer.nas that ignores the "valid" flag for the latest pilot that has joined. Please check if this solves the problem. Torsten -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july ___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
[Flightgear-devel] pilot list error
As I mentioned before , I've been looking for the missing player in the pilot list and model view. If I comment out the 2 lines below "foreach" , in multiplayer.nas at line 374 ... foreach (var n; props.globals.getNode("ai/models", 1).getChildren("multiplayer")) { #if (!n.getNode("valid", 1).getValue()) # continue; ... I then get the correct in /ai/models, and I can now step through ALL the players in "model view" , and they all appear in the pilot list. I don't consider this a fix since I dont know why the above lines would cause a problem,but it's a start. Cheers -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
Re: [Flightgear-devel] pilot list...
I've mentioned the problem here before, but as far as I know, no one has pinned it down, because all of the Win32 builds from June forward exhibit this behavior. Thanks for looking into it! -R. Robert M. Shearman, Jr. Transit Operations Supervisor, University of Maryland Department of Transportation also known as rm...@umd.edu From: syd adams To: flightgear-dev Sent: Sunday, August 30, 2009 1:24:13 AM Subject: [Flightgear-devel] pilot list... While attempting to hunt down a chat message problem , (no messages displayed from certain players), I found that the pilot list is always 1 player short , and the "model view" skips that player. They all appear in ai/models , but num-players is always show 1 less player. I've been sifting through the multiplay.nas , but thought I'd mention it here in case someone has already looked at the problem . cheers -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel
[Flightgear-devel] pilot list...
While attempting to hunt down a chat message problem , (no messages displayed from certain players), I found that the pilot list is always 1 player short , and the "model view" skips that player. They all appear in ai/models , but num-players is always show 1 less player. I've been sifting through the multiplay.nas , but thought I'd mention it here in case someone has already looked at the problem . cheers -- Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july___ Flightgear-devel mailing list Flightgear-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/flightgear-devel