1.1.2 Magic Smoke Industries Infernal Robotics 2.0.2

[Ziw]

39 minutes ago, TrHollywood said:

No what i'm talking about is the parts don't like to detach from the host ship your moving them from. They snap where you move them but then the game studders for a second and they move right back to where they were also sending my ship either flying off in to the dark abyss or creating a pretty light show. 

Oh, that is new  What if you just drop them first and then pick up and attach to the new ship? That should definitely work, at least it worked before when I tested it.

 

[maculator]

It's not really a problem but it's something that made me curious:

Spoiler

This is the version makes the parts show only in the IR-category:

	@PART[*]:HAS[@MODULE[MuMechToggle]]
	{
		%category = none
	}

However my first attempt was this one:

	@PART[*]:HAS[@MODULE[MuMechToggle]]
	{
		!category = DELETE
	}

Well it didn't work as I thought it does; It put the parts inside the Fuel-Category. I don't know how this is possible but maybe someone here knows what happend. The MM-patch seems to be correct.

Edit: I also have a real question I hope someone can help me with: The joints seem to be verry weak, even KJR doesn't change that. This makes my dream of a plane with movable wings pretty hard to achieve. Is there a way to force those mods to work together or can I edit the configs/write a MM-patch?.

Thanks

Edited February 29, 2016 by maculator
Trying to write correct english :( 2.0

[Ziw]

3 hours ago, maculator said:

 

Edit: I also have a real question I hope someone can help me with: The joints seem to be verry weak, even KJR doesn't change that. This makes my dream of a plane with movable wings pretty hard to achieve. Is there a way to force those mods to work together or can I edit the configs/write a MM-patch?.

Thanks

For now you can only achieve this using Quantum Struts or Active Struts to handle the weight and lifting power of the wings. Alternatively you can scale up the hinge/rotatron parts in part.cfg to make them bigger with bigger nodes and bigger mass - it has been reported to help as well. 

[Guernica]

I'm trying to use the Free Docking Washer to make a rotating portion of a spacecraft, but nothing will allow it to spin. I've made sure it is not connected to a docking port, but it still acts as if it is locked in place when I fire rockets to spin the section up.

[Ziw]

15 minutes ago, Guernica said:

I'm trying to use the Free Docking Washer to make a rotating portion of a spacecraft, but nothing will allow it to spin. I've made sure it is not connected to a docking port, but it still acts as if it is locked in place when I fire rockets to spin the section up.

Did you try an alternative free rotating part from Zodius Model Rework (see below)?

 

[Guernica]

So I did a bit more investigating and testing and the problem seems to be with TweakScale. Using the default size of 1.25m, it rotates fine, but if I scale it up at all, it becomes frozen. Further research suggest this has been a problem going back quite awhile (present in .24.5 of the game).

I tried the model re-work, but it does not offer a part that is flat so-to-speak. The bearing rotatron just does not look good on a station where you'd expect the crew to transfer between rotating and non-rotating parts, even scale to Large+.

I feel like a spoiled brat asking for more milk.

[V8jester]

4 minutes ago, Guernica said:

So I did a bit more investigating and testing and the problem seems to be with TweakScale. Using the default size of 1.25m, it rotates fine, but if I scale it up at all, it becomes frozen. Further research suggest this has been a problem going back quite awhile (present in .24.5 of the game).

I tried the model re-work, but it does not offer a part that is flat so-to-speak. The bearing rotatron just does not look good on a station where you'd expect the crew to transfer between rotating and non-rotating parts, even scale to Large+.

I feel like a spoiled brat asking for more milk.

If you must have huge docking washers. It can be done but involves copying the cfg file for the docking washer and editing the copy. This is done in the folder "GameData/MagicSmokeIndustries/Parts/Legacy/dockingwasher" Just right click, copy and edit the copy right there. This will provide you with a new part in KSP but you must change the name and title or it wont work. I suggest notepad++

PART
{
name = dockingwasher_stdScaleable = Change name to whatever you'd like

rescaleFactor = 1.25 = Change rescaleFactor = 1.25 - to something like 2.5 or bigger

title = Docking Washer Standard = Change Title to whatever you want

 

[Ziw]

1 hour ago, Guernica said:

35 minutes ago, Guernica said:

So I did a bit more investigating and testing and the problem seems to be with TweakScale. Using the default size of 1.25m, it rotates fine, but if I scale it up at all, it becomes frozen. Further research suggest this has been a problem going back quite awhile (present in .24.5 of the game).

I tried the model re-work, but it does not offer a part that is flat so-to-speak. The bearing rotatron just does not look good on a station where you'd expect the crew to transfer between rotating and non-rotating parts, even scale to Large+.

I feel like a spoiled brat asking for more milk

 

It's a bit old, but it should work nonetheless.

Edited February 29, 2016 by Ziw

[allmhuran]

On 2/29/2016 at 9:17 PM, maculator said:

Edit: I also have a real question I hope someone can help me with: The joints seem to be verry weak, even KJR doesn't change that. This makes my dream of a plane with movable wings pretty hard to achieve. Is there a way to force those mods to work together or can I edit the configs/write a MM-patch?.

I struggled with this for a long time and I have something that might help.

It's true that there is a lot of weakness in the pivoting physics joints in IR parts: they will rotate even when you don't want them to when subject to sufficient force. But that's compounded by the connection sometimes being weak (floppy) between the IR part and the part to which it's attached. This is actually a problem with all low mass parts in KSP. Based on my re-testing tonight It seems to have been improved from 1.04 to 1.05 (I just installed 1.05 yesterday), or perhaps it's only certain parts that do it.

In any case, the second problem can be alleviated by connecting a physicsless part on each side of the IR part. For example (fuel tank)-(cubic octagonal strut)-(IR joint)-(cubic octagonal strut)-(fuel tank). This will at least make the joint between the IR part and its neighbours rigid.

 

[allmhuran]

KOS 0.19.0 has just been released: https://www.reddit.com/r/Kos/comments/48klw7/new_release_v0190/
 

The KOS release notes themselves don't mention anything specific about changes in IR other than a general bugfix ("IR:AVAILABLE was reporting incorrectly"). Any word on that? I'm particularly interested in the "KOS CPU's on remote (unfocused) vessels being able to control servos on that vessel" thing for the sake of my AI mechs. 

Edit: Looks like it's all good, since the IR documentation page now has descriptions like this: " Lists all Servo Groups for the Vessel on which the script is being executed. On IR versions prior to 0.21.5 will always return servo groups for current focused vessel " Yippee!

 
 

Edited March 2, 2016 by allmhuran

[Ziw]

21 minutes ago, allmhuran said:

KOS 0.19.0 has just been released: https://www.reddit.com/r/Kos/comments/48klw7/new_release_v0190/
 

The KOS release notes themselves don't mention anything specific about changes in IR other than a general bugfix ("IR:AVAILABLE was reporting incorrectly"). Any word on that? I'm particularly interested in the "KOS CPU's on remote (unfocused) vessels being able to control servos on that vessel" thing for the sake of my AI mechs. 

Edit: Looks like it's all good, since the IR documentation page now has descriptions like this: " Lists all Servo Groups for the Vessel on which the script is being executed. On IR versions prior to 0.21.5 will always return servo groups for current focused vessel " Yippee!

 
 

You might want to use this dll in your IR to use the new integration properly.

https://github.com/MagicSmokeIndustries/InfernalRobotics/releases/tag/0.21.5-pre

 

I will not make it a proper release though, I have bigger plans for the next update.

[allmhuran]

Awesome, cheers! I will use it for some experimentation <lightning in background, maniacal laughter, etc>

[allmhuran]

Might be a bit of a bug with the API. I made a simple little program to test controlling the servo with axis input. It seemed to work, but then I tweaked a couple of values (just set the speed to be the square of the input) and the servos stopped responding to the command. Setting the speed back to the original didn't fix it, nor did reverting to launch.

Here's the program code: http://pastebin.com/nC5C6Lx7

All you need to use the program is a vehicle with a servo tagged as "waist".

And here's a quick gfy showing the problem. Notice that the program itself seems to be working, the print statement on line 8 is spitting out values as expected.

 

 

Edited March 3, 2016 by allmhuran

[Rushligh]

I've had similar problems without using scripts. as of recently i was using pistons attached to KF tracks to make a variable height platform (for launching submarines and boats) and noticed that as soon as i applied piston movement though action groups some of the pistons wouldn't respond at all. the other would jitter and shake, only moving half their supposed distance.

I believe it may have to do with Kerbal Joint Reinforcement but im not certain.

[Ziw]

1 hour ago, allmhuran said:

Might be a bit of a bug with the API. I made a simple little program to test controlling the servo with axis input. It seemed to work, but then I tweaked a couple of values (just set the speed to be the square of the input) and the servos stopped responding to the command. Setting the speed back to the original didn't fix it, nor did reverting to launch.

Here's the program code: http://pastebin.com/nC5C6Lx7
 

You said reverting to launch did not help? what if you scrap it and launch once again?

Please do 2 things for me: test with 0.21.4 release dll just to rule out the simplest possibility (though 99% it won't help) and give me the log file (KSP.log should be sufficient)

 

[allmhuran]

4 hours ago, Ziw said:

You said reverting to launch did not help? what if you scrap it and launch once again?

Please do 2 things for me: test with 0.21.4 release dll just to rule out the simplest possibility (though 99% it won't help) and give me the log file (KSP.log should be sufficient)

 

I shall do so at my earliest convenience (in about 6 hours when I get home from work). 

It seemed like it was somehow, sometimes referencing a ghost servo. While "debugging" the problem I executed the same logic to find the servo, but just put it into a terminal variable. I then manually executed things like

set waist:speed to 10.
print waist:speed.

and, as I recall, the printed speed wasn't 10 (I will confirm this tonight). But it didn't throw an error about an unknown suffix "speed", so it was referring to some servo or other. It sounds like you already suspect something like this based on what you said in your first paragraph. Perhaps there's a memory leak somewhere with servos not being cleaned up?

Edited March 4, 2016 by allmhuran

[allmhuran]

4 hours ago, Rushligh said:

I've had similar problems without using scripts. as of recently i was using pistons attached to KF tracks to make a variable height platform (for launching submarines and boats) and noticed that as soon as i applied piston movement though action groups some of the pistons wouldn't respond at all. the other would jitter and shake, only moving half their supposed distance.

I believe it may have to do with Kerbal Joint Reinforcement but im not certain.

This is probably an issue with the construction of the vessel. If you have the parts turned the *wrong way around*, or if you have stray struts, or any of a few other things that I haven't been able to identify, your servos may behave strangely.

* The right way around is to have the side of the servo with the yellow border and arrows attached to the part of your craft containing the root part

Edited March 4, 2016 by allmhuran

[Ziw]

6 hours ago, allmhuran said:

It seemed like it was somehow, sometimes referencing a ghost servo. While "debugging" the problem I executed the same logic to find the servo, but just put it into a terminal variable. I then manually executed things like

set waist:speed to 10.
print waist:speed.

and, as I recall, the printed speed wasn't 10 (I will confirm this tonight). But it didn't throw an error about an unknown suffix "speed", so it was referring to some servo or other. It sounds like you already suspect something like this based on what you said in your first paragraph. Perhaps there's a memory leak somewhere with servos not being cleaned up?

Yes, that is my concern. What puzzles me most is the absence of kOS errors if its the case. That is why I want the log, maybe there is something there.

[allmhuran]

OK, reporting back:

Using 0.21.5-pre (I will retest with 0.21.4 in a short while):

• If I run KSP, build a new craft, and launch it, it seems to work (3 worked out of 3 attempts)
• If I revert a craft to launch, it seems to fail (0 worked / 3 retries).
• If I revert to SPH and relaunch it seems to fail (0/3).
• If I revert to SPH, click to load a craft, pick the same craft, and relaunch, it seems to fail (0/3).
• If I revert to SPH, click to load a craft, pick a different craft, then load back to the original craft, then relaunch,  it seems to fail (0/1).
• If I revert to SPH, click to build a new craft, set a root piece, merge in the old craft, and relaunch.... it fails (0/1)

I ran a few other tests but it started getting convoluted

Here is a short video of everything that I did during one restart of the game.

The log file which you can find here: https://drive.google.com/file/d/0B2m8Wm1qifJudWIwTnVJdWpWV2M/view?usp=sharing was generated by what you see in the video.

The program I used is slightly simpler than the previous one and can be found here: http://pastebin.com/g2dd4arV

The quality of the video is reduced and it is sped up in order to keep the upload size down, upload bandwidth in Australia is terrible and I wanted to get the results up earlier than (several hours from now).

 

Edited March 4, 2016 by allmhuran

[Ziw]

All this indicates a problem with servo object pointing incorrectly somewhere inside kOS integration

I feel like 0.21.4 dll will work the same, so I need to look deep into code. Will revert as soon as I find out  anything.

[allmhuran]

Yep confirmed in 0.21.4. On my first attempt it worked, when I reverted to launch the program ran but there was no response from the servo. Weird that it's only on some occasions that I get those extra errors (like ":ismoving not found" )

[Ziw]

15 minutes ago, allmhuran said:

Yep confirmed in 0.21.4. On my first attempt it worked, when I reverted to launch the program ran but there was no response from the servo. Weird that it's only on some occasions that I get those extra errors (like ":ismoving not found" )

There were so many changes to kOS since my PR was submitted... I confirm that I have reproduced the bug and it's not IR bug, because IR UI still moves the part properly. Looking into integration code.

 

[Ziw]

3 hours ago, allmhuran said:

Yep confirmed in 0.21.4. On my first attempt it worked, when I reverted to launch the program ran but there was no response from the servo. Weird that it's only on some occasions that I get those extra errors (like ":ismoving not found" )

Good news, with the help from other kOS devs I've found the fix. It will be in the next kOS update.

[allmhuran]

That was fast. Again.

[Rushligh]

17 hours ago, allmhuran said:

This is probably an issue with the construction of the vessel. If you have the parts turned the *wrong way around*, or if you have stray struts, or any of a few other things that I haven't been able to identify, your servos may behave strangely.

* The right way around is to have the side of the servo with the yellow border and arrows attached to the part of your craft containing the root part

I've tried reversing them to no avail. when i do the opposite side of my craft moves shakily and the other side doesn't respond. I'm using NeatFuture Construction for the chassis.

I figured out the problem. it has to do with weight, tweak scale, and symmetry. NTF construction hubs wont use symmetry (IE you have to place left and right separately). this seems to do something to the unity physics making the two sides not treated the same. the reason why my piston problem changed sides it because i changed which side i made first when i reversed the pistons.

as for the jittering/not responding it seems to be because i overloaded them with weight. (6 pistons holding 32 tonnes of submarine). Tweak scale doesn't seem to proportionately chance how much weight they can handle, just the size and electric draw.

I rebuilt the chassis with stock beams using symmetry and with 12 pistons on each side and it worked for lighter craft. perhaps there's a mod or a way to make IR joints handle greater weight/torque?