Test of continuing KER developments [0.24.2]

[JWag]

I ran into something like Boomerang's issue, but I didn't document whatever I did to fix it.

I was working with an old .23.5 career in .24 and had added KER and Padishar's DLLs (the career had been completely stock, but I had progressed a few nodes up the tech tree). I also added a ModuleManager script from one of these threads that automatically attached a KER part to all of the command parts so I wouldn't have to add the parts in the VAB.

When editing one of my old ships (actually just a Mk 1 capsule) in the VAB , KER wasn't showing up in either toolbar. On the launch pad, I had a similar KER menu as the images Boomerang posted. I'd see some data in the Surface interface, but it would clear out when physics started and I'd have NREs all up in my logs.

Turns out I forgot to go back into the science tree and research the KER components. After I did that and recovered the pod from the pad and cleared the VAB, everything worked as expected.

Not sure if that'll help or confuse things further, but I figured I'd throw it out there.

[Padishar]

Turns out I forgot to go back into the science tree and research the KER components. After I did that and recovered the pod from the pad and cleared the VAB, everything worked as expected.

Well, you should only need to research the KER parts if you actually use them so if you use an MM config to add the KER modules to command pods and don't have an actual engineer part on the ship then it should still work fine. Without the output_log.txt there isn't much I can do about this one...

Testing now, I'll update with any changes or at least a new log in a few minutes.

Edit: No change in behavior and physics loading took long enough that I was able to scan the 'Surface' tab and everything looked the way it ought to and certainly no "-OVF-". Here's the latest log: http://justpaste.it/gb6c

Doh, sorry, it looks like the zip I uploaded contained the wrong DLLs. Can you try getting the zip again (the modified date of the two DLLs should be 20/07/2014 19:41 GMT) and giving it another go. It certainly shouldn't be giving an array index exception in that function anymore...

Edited July 20, 2014 by Padishar

[Boomerang]

Well, you should only need to research the KER parts if you actually use them so if you use an MM config to add the KER modules to command pods and don't have an actual engineer part on the ship then it should still work fine. Without the output_log.txt there isn't much I can do about this one...

Doh, sorry, it looks like the zip I uploaded contained the wrong DLLs. Can you try getting the zip again (the modified date of the two DLLs should be 20/07/2014 19:41 GMT) and giving it another go. It certainly shouldn't be giving an array index exception in that function anymore...

We have success! This time the Surfaces menu stayed open (and I could open/close it normally)!

Here's the log: http://justpaste.it/gb6j

And it was the Terminal Velocity giving the '-OVF-' error, which makes me think that it's because I'm running NEAR? My computer's about to run out of charge, but I'll try running without NEAR in a little while and get back to you shortly. Thanks for all the work!

Edited July 20, 2014 by Boomerang

[Padishar]

We have success! This time the Surfaces menu stayed open (and I could open/close it normally)!

Here's the log: http://justpaste.it/gb6j

And it was the Terminal Velocity giving the '-OVF-' error, which makes me think that it's because I'm running NEAR? My computer's about to run out of charge, but I'll try running without NEAR in a little while and get back to you shortly. Thanks for all the work!

Good, that makes more sense. KER disables the 5 atmospheric readouts when FAR is installed and it looks like it needs to do the same for NEAR too. I'm off to bed in a minute but I'll try to take a look at this tomorrow...

[Boomerang]

Good, that makes more sense. KER disables the 5 atmospheric readouts when FAR is installed and it looks like it needs to do the same for NEAR too. I'm off to bed in a minute but I'll try to take a look at this tomorrow...

Sounds great and I can confirm that removing NEAR stopped KER from throwing the -OVF- error. Everything seems kosher that way. I'll reinstall NEAR since the last patch kept KER from breaking and just keep an eye out for the next update. Thank you very much for all the time and patience, this has been one of the best troubleshooting experiences I've had.

[Padishar]

I have updated the zip linked in the first post to include a check for NEAR that disables the atmospheric readouts in the same way as the existing FAR check.

[mic_e]

Padishar: Just to be clear on the issue: Did you "take over" KER development? cybutek seems to have fallen back into "maintenance mode" a while ago, and stopped accepting my pull requests (such as Biome/Impact Biome, which I've noticed you've implemented).

I've also written code that calculates FAR drag, leading to correct terminal velocities / stock-equivalent mass-specific drag coefficients. You should take a look at it, update it to .24 and merge it . The code is Reflection-based, so it works for both stock and FAR setups (see lines 465, 631 and 664): flightengineer.cs

[sarbian]

I've also written code that calculates FAR drag, leading to correct terminal velocities / stock-equivalent mass-specific drag coefficients. You should take a look at it, update it to .24 and merge it . The code is Reflection-based, so it works for both stock and FAR setups (see lines 465, 631 and 664): flightengineer.cs

I think I ll have a talk with you about that FAR code when I m less busy with .24 stuff mic_e. That FAR code could be quite usefull ^^

[Padishar]

Padishar: Just to be clear on the issue: Did you "take over" KER development? cybutek seems to have fallen back into "maintenance mode" a while ago, and stopped accepting my pull requests (such as Biome/Impact Biome, which I've noticed you've implemented).

I've also written code that calculates FAR drag, leading to correct terminal velocities / stock-equivalent mass-specific drag coefficients. You should take a look at it, update it to .24 and merge it . The code is Reflection-based, so it works for both stock and FAR setups (see lines 465, 631 and 664): flightengineer.cs

Originally, I offered to try fix the various problems in the simulation code because one of the stations I was building caused it to lock up completely and when I looked at the code it was fairly obvious where the biggest issues were and what needed to be done. Since then, most of the changes I have done have been to fix issues with the simulation code though I've also added a few other features and fixed other bugs. Cybutek has been working on a significant rewrite of the mod's general framework and UI code and is quite close to an alpha release.

I have seen your pull request on the main repo and wondered why it was never integrated as it looks pretty good to me. I've just had to disable the same readouts for NEAR (simpler version of FAR) so it would be very useful to have similar code for that too. Hopefully Cybutek will see this and reconsider merging it, though it might be helpful if you could merge from his latest code again and tidy it up to make it easier for him...

[AlienCollective]

I'm having the issue with the Engineer window not appearing. This happens with both 32- and 64-bit versions. I'm using 0.6.2.7, and have the same behaviour with and without your dev DLLs. Screencast:

https://dl.dropboxusercontent.com/u/68829/engineer1.mp4

Engineer is the only installed mod. I even tried removing NASAmission.

Not sure if it matters, but my system:

Windows 7 SP1 x64

AMD A10-5800K APU 3.8 GHz

16 GB DDR3 1866 MHz

Radeon HD 7870

Two other points: KSP 0.24 has been causing my computer to crash, as in full power off as if someone had yanked the power cable. May or may not be related, may or may not be a result of Engineer. I haven't run that one down yet. Second, I have occasionally been able to get Engineer working (somehow), but it usually behaves like in the screencast.

Please let me know if you need any more info.

UPDATE: Deleting the mod folder and reinstalling worked. Maybe a config conflict?

Edited July 23, 2014 by AlienCollective

[BudgetHedgehog]

I'm having the issue with the Engineer window not appearing. This happens with both 32- and 64-bit versions. I'm using 0.6.2.7, and have the same behaviour with and without your dev DLLs.

This isn't mine, but this is an output log from someone who told me they have the same problem (including the disappearing ship name field). They also apparently had the latest versions of KER: https://www.dropbox.com/s/ui36klo1w8mtpur/output_log.txt

It's from this Reddit thread.

EDIT: Found the error in question, I think:

Exception in BuildEng.OnLoad: System.IndexOutOfRangeException: Array index is out of range.

  at Engineer.Settings.Load (System.String filename, .Vessel vessel) [0x00000] in <filename unknown>:0 

  at Engineer.BuildEngineer.OnLoad (.ConfigNode node) [0x00000] in <filename unknown>:0 

(Filename: C:/BuildAgent/work/d63dfc6385190b60/artifacts/StandalonePlayerGenerated/UnityEngineDebug.cpp Line: 49)

stage count is: 0

(Filename: C:/BuildAgent/work/d63dfc6385190b60/artifacts/StandalonePlayerGenerated/UnityEngineDebug.cpp Line: 49)

BuildEngineer: OnSave

(Filename: C:/BuildAgent/work/d63dfc6385190b60/artifacts/StandalonePlayerGenerated/UnityEngineDebug.cpp Line: 49)

Exception in BuildEng.OnSave: System.IndexOutOfRangeException: Array index is out of range.

  at Engineer.Settings.Load (System.String filename, .Vessel vessel) [0x00000] in <filename unknown>:0 

  at Engineer.Settings.Save (System.String filename, .Vessel vessel) [0x00000] in <filename unknown>:0 

  at Engineer.Settings.Set[String] (System.String key, System.String value) [0x00000] in <filename unknown>:0 

  at Engineer.Settings.Set (System.String key, System.String value) [0x00000] in <filename unknown>:0 

  at Engineer.BuildEngineer.OnSave (.ConfigNode node) [0x00000] in <filename unknown>:0 

Edited July 23, 2014 by ObsessedWithKSP

[Master Tao]

Please let me know if you need any more info.

Even if ObsessedWithKSP's info is the same problem, your log files would still help, although the video is interesting. Never seen the vessel name display disappear before.

Two other points: KSP 0.24 has been causing my computer to crash, as in full power off as if someone had yanked the power cable.

That sounds like the CPU is overheating, or a serious driver issue. Find a program to monitor your CPU temperature, make sure the fan isn't clogged with dust, and update your Radeon drivers and DirectX.

[Padishar]

Well, an output_log.txt from a run of KSP doing the same as in your screencast might be helpful.

Check in the GameData/Engineer/PluginData/Engineer folder. If there is a build_engineer.cfg (and/or flight_engineer.cfg) file that is empty then delete it. It appears that sometimes when KSP crashes these files can be left present but empty, which seems to confuse KER...

Doh! Double ninja'd... Try checking the .cfg files like I suggested. I'll try to take a look at the config loading and saving code in the next day or two...

Edited July 23, 2014 by Padishar

[mic_e]

I think I ll have a talk with you about that FAR code when I m less busy with .24 stuff mic_e. That FAR code could be quite usefull ^^

I'm usually idling in #kspmodders and #kspofficial, just ping me and I _might_ be there.

...

I see. KER2 seemes to have been in development, and KER development halted, for over a year or so now, so I was wondeering whether it was still happening. It's just getting very frustrating to see your code not getting merged with little feedback, and having to manually do it myself after every update... In fact, it got so frustrating that I did a proof-of-concept readout plugin myself (designed to allow adding readouts from external DLLs, currently just displays Pe/Ap/v_escape)... Edited July 23, 2014 by mic_e

[BudgetHedgehog]

Check in the GameData/Engineer/PluginData/Engineer folder. If there is a build_engineer.cfg (and/or flight_engineer.cfg) file that is empty then delete it. It appears that sometimes when KSP crashes these files can be left present but empty, which seems to confuse KER...

Just so you know, I've been told that fixed the problem. The dude/dudette says thanks

[AlienCollective]

Even if ObsessedWithKSP's info is the same problem, your log files would still help, although the video is interesting. Never seen the vessel name display disappear before.

Here's my logfile, with the config files from the offending version of Engineer thrown in for good measure.

https://dl.dropboxusercontent.com/u/68829/engineer-logs.zip

That sounds like the CPU is overheating, or a serious driver issue. Find a program to monitor your CPU temperature, make sure the fan isn't clogged with dust, and update your Radeon drivers and DirectX.

I wasn't actually trying to hijack your thread with my computer problems, but thanks for the tips.

[Padishar]

Here's my logfile, with the config files from the offending version of Engineer thrown in for good measure.

Yes, the build_engineer.cfg file is empty as I suggested. Deleting that should have been enough to fix the issue rather than the more drastic full reinstall...

[AlienCollective]

Yes, the build_engineer.cfg file is empty as I suggested. Deleting that should have been enough to fix the issue rather than the more drastic full reinstall...

It seems to be recreating the file and regressing after a few runs. I'll let you know when I have steps to reproduce.

[Padishar]

It seems to be recreating the file and regressing after a few runs. I'll let you know when I have steps to reproduce.

I have just updated the zip linked in the first post to include a fix for this issue. It will log any malformed lines in the settings config file and will also log if nothing useful is loaded from the file. This eliminates the exceptions that cause the various bad things to happen (mainly the window code failing and it not saving out a correct settings file).

[Padishar]

I have just updated the zip linked in the first post. It includes various fixes to the 0.6.2.8 version and also makes the tweakables controllable via the settings dialog rather than by using replacement PartModules. See the change log in the first post for details.

[BudgetHedgehog]

Good stuff. Only trouble is, I'm running KER 1.0.0.1

[Padishar]

Good stuff. Only trouble is, I'm running KER 1.0.0.1

Yeah, I'm hoping that Cybutek will be able to merge my pull request into 0.6.x.x and also include any of the fixes that apply to 1.x.x.x fairly soon...

[Padishar]

Yeah, I'm hoping that Cybutek will be able to merge my pull request into 0.6.x.x and also include any of the fixes that apply to 1.x.x.x fairly soon...

KER 0.6.2.9 has been released including my changes so there is no need to install this version until I update it with new changes based on the 0.6.2.9 code.

[Padishar]

I have now released a new dev version based on 0.6.2.9 that should fix various of the issues when empty stages or stages with just launch clamps cause the deltaV calculations to go wrong.

I would appreciate a few people giving it a test before it gets rolled into the next official version.

I am also working on fixing the issues when the root part of the craft is decoupled. Hopefully I will get this done in the next day or two...

[Padishar]

I have just released a new dev version with a fix for a bug reported over in the main KER thread which causes the simulation code to fail with an "invalid parameter" exception.