Need Help Stat, clients getting weird Error.

[Pryme8]

So I delivered a update to one of my clients today, and they are getting this error while I am not:


The scene compiles fine for me and runs great,  I have no clue wtf or how to ever start debugging this especially since I can not recreate it.

I assume they are using all the same files I served to them that I am running on my local, and I am hung over as all get out today so my brain is not working... and am clueless on this.

@BitOfGold Cannon god please come help. >_<

[adam]

My guess is that you are dealing with a timing issue.  It looks like the physics plugin is trying to access the mesh before it's ready.

[Pryme8]

couldn't be, all meshes are preloaded and the physics does not attach until the whole scene and is constructed.  I did think that initially as well, but Im not able to get Three different PC to drop the error all with way different specs, and nope.

This is super odd.

[adam]

Did you tell them to delete their cache?

[Pryme8]

>_< I really hope that is it.

[Pryme8]

nope, and I did not think it was cacheing because he made a fresh directory on his local host

[Pryme8]

Yeah and the weird part is he remoted it to anouther PC and got the same error?!  And I have yet to be able to produce it on 3 of my machines so I'm stumped

[dbawel]

send me a link to the scene. This is solvable.

[Pryme8]

I will send you an email now dave

[Pryme8]

So the plot thickens, I could not recreate the error at all, until I booted it up in firefox when I do that I get the same error as my client but at a different position.  The craziness is if I load it into chrome no error.  I need to pull put the big guns on this one and am willing to pay someone if I am unable to figure it out (which is odd). 

This has to be fixed today or AIS Goverment will lose all confidence in me I have a feeling.

[Pryme8]

Ok one good thing in my favor now, it works with no error on his colleges computer so that means we are having different systems handle the scene completely differently.  Umm timing issue?  Maybe set the timouts to like 30 instead of 0?

This is really really weird.  Oh and the crazyness is I changed where a console.log was and nothing else and then the error disappeared then 10 mins later came back when I made no other changes (and no its not a cache issue I do not ever work with a cache on).  I have never seen behavior like this in my life.

[GameMonetize]

can you try to use babylon.max.js to see where this is happening?

Is there a way to get a link to the repro case?

[Pryme8]

@Temechon or @RaananW paid debug? PM me!

Um I think i have narrowed it down to cannon trying to make in imposter on a mesh that has no vertex data, which is odd I'm going to check my obj files to make sure this is not the case but I'm kinda intrigued why half the clients do not receive this error... if its an error its and error It should not compile on anyone's.  And until i served the product had no clue there was even a problem.

Due to it being a product not technically owned by me I do not think I could repo it without their permission but I will ask and see if they are cool with that.

[Pryme8]

I think I may have found a work around!

ill post if I got it, I found a Null Object was the parent object for one of my mesh groups.  I am going to fix that and see if that stops the error.

[RaananW]

Yeah, null parent might cause this, as the physics engine counts on parenting when calculating the physics objects.

If you want to show me a demo of what happens, I might be able to help further.

[adam]

@Pryme8 Can you let us know how this turned out?  Thanks

[Pryme8]

I was never able to figure this out and trying to fix it only compounded the issue now.  I have found now that c4d r17 and c4d r14 export obj files in a complete different hierarchy and stick them in a null object no matter what you do with r17 (eajklfhsdjklahsfhjklasdfashklf bullshit).  So I have to make sure I do all my preset exports for my system with r14.  Now the odd part is that Chrome does not skip a beat with this parent null.

And now for the worse news... I was just informed last night that due to the clients requirements (because they are a federal entity) of what browser they have on their employees computers I have to make this work on IE11... and I about dropped a bollocks...

A paid debug is still in the air for one of you guys who think you can fix it.  PM me if you are feeling froggy.

[adam]

I primarily use IE11 + Visual Studio when developing projects in BJS.  

[Pryme8]

Ok so there is hope?

[adam]

Definitely

[RaananW]

Babylon is compatible with IE11. It should work.

Do you have a link to the scene? if it is private, PM me, I just want to see what actually happens.

[Pryme8]

so besides troubleshoot this error, it looks like Im going to inject Chrome into IE11 right on the page and just overwrite all the webgl stuff...  pretty much going to have the webpage gut IE and replace it with a Chrome brain.  I am under certain specific obligations due to it being a Federal Contract, and IE11 is now one of them, which just doing research on its GL support yeah its there but its experimental and limited.  Im not aloud to make an executable, and everything has to be hosted int he cloud so yeah im sorry IE but im about to perform surgery.

[adam]

I'd at least try it in IE11 before dismissing it.

[adam]

If you are talking about Chrome Frame, it is no longer supported.

https://www.chromium.org/developers/how-tos/chrome-frame-getting-started

[Pryme8]

arg yeah I was and that's a disappointment....



Ok just a random thought.... what if I turn all of the functions that chrome has built in to it for webGL and serialize them and then overwrite the other browers methods? is that possible?

Ill try to give IE11 a shot but the thing im worried about is performance point blank and what I have seen outa IE seems lacking.  Also it seems hindered by the same odd error the firefox has with it that chrome skips over without a hickup.