Youen

[old thread] Trajectories : atmospheric predictions

Recommended Posts

2 hours ago, linuxgurugamer said:

I was trying to have the Suicide Burn mod work on vessels which aren't the active vessel;  ie:  dropped probe, etc.

Sounds interesting.

I took a look in the code and I see one main problem:

In order to calculate the forces on a ship (see StockAeroUtil), we interate over parts and query variables like bodyLiftMultipilier and accesses the drag cube data of a part. What I don't know is that if a non-active vessel object actually has all these variables initialized. My guess is that it is not.Take a look at StockAeroUtil.SimAeroForce, around line 160. If these variables are initialized for a non-active vessel, simulation should be doable. If not, then no dice.

If you want this to be compatible with FAR, you will have to check that as well.

Either way, this would be quite a big refactoring that I don't have time for: you would have to do the heavy lifting. If you want it merged, then you would also have to make sure that nothing (new) breaks for the main use case.

 

Share this post


Link to post
Share on other sites
2 minutes ago, Kobymaru said:

Sounds interesting.

I took a look in the code and I see one main problem:

In order to calculate the forces on a ship (see StockAeroUtil), we interate over parts and query variables like bodyLiftMultipilier and accesses the drag cube data of a part. What I don't know is that if a non-active vessel object actually has all these variables initialized. My guess is that it is not.Take a look at StockAeroUtil.SimAeroForce, around line 160. If these variables are initialized for a non-active vessel, simulation should be doable. If not, then no dice.

If you want this to be compatible with FAR, you will have to check that as well.

Either way, this would be quite a big refactoring that I don't have time for: you would have to do the heavy lifting. If you want it merged, then you would also have to make sure that nothing (new) breaks for the main use case.

 

I was afraid of that.

I don't have the time now, so will have to put this on the side for now.

Thanks for looking

Share this post


Link to post
Share on other sites

@Kobymaru

Do you know when you will be doing the next release?  I'm heading into the home stretch on the Suicide Burn mod, and will need this.

If its not going to be for a while, I can always package a local version of it until you do get it out

Let me know, thanks

Share this post


Link to post
Share on other sites
5 hours ago, linuxgurugamer said:

Do you know when you will be doing the next release?  I'm heading into the home stretch on the Suicide Burn mod, and will need this.

The current plan is to release it when KSP 1.3.0 drops.

Share this post


Link to post
Share on other sites

Ok, this is the fastest turnaround from "not possible/not to be implemented" to "here ya go!" I've ever seen

Share this post


Link to post
Share on other sites

Okay, have a problem with this mod...

If you're in map view, with trajectories visible, and you stage (either via MechJeb autostaging or by Smart Parts), the game WILL lock up and crash.

Share this post


Link to post
Share on other sites

So.. Does it work with 1.2.2 FAR stable release?  

Because trajectories had problem with new FAR release before.

Share this post


Link to post
Share on other sites

just tried the mod with the upper stage of the kerbal X reentering, appears stock atmosphere trajectory is working...

EDIT: just had a look at the "one Mk1 pod" quicksave, that appears to still be wildly inaccurate

Edited by The Crusher 862

Share this post


Link to post
Share on other sites
On 5/27/2017 at 6:20 AM, ssd21345 said:

So.. Does it work with 1.2.2 FAR stable release?  

Because trajectories had problem with new FAR release before.

It works excellently

Share this post


Link to post
Share on other sites
On 25.5.2017 at 4:17 AM, MaverickSawyer said:

Okay, have a problem with this mod...

If you're in map view, with trajectories visible, and you stage (either via MechJeb autostaging or by Smart Parts), the game WILL lock up and crash.

Please narrow down your Mods to the minimal combination where this bug still occurs, check if it occurs for every vessel in every situation (or just specific ones) and then post logs, preferrably in a pastebin somewhere.

 

On 29.5.2017 at 0:49 AM, The Crusher 862 said:

EDIT: just had a look at the "one Mk1 pod" quicksave, that appears to still be wildly inaccurate

Yeah, that's a problem. I tried to use Trajectories for an Eve-Atmospheric entry, and the prediction was so wrong that it was completely useless. Puppies cried tears that day.

This is a tough one to fix, but help is of course appreciated. I think I'll proceed to write some sort of Telemetry/Logging module, because I need to "see" the numerical values to start to figure out what's wrong.

 

On 24.5.2017 at 8:17 PM, dlrk said:

Ok, this is the fastest turnaround from "not possible/not to be implemented" to "here ya go!" I've ever seen

Well I always said "not to be implemented unless I get pull requests". And while DennyTX was not willing to contribute to Trajectories directly, he did set a nice (almost copy-pastable) code example with his Landing Aim mod. And that's what happens when I get code examples :wink: 

But even so, I'm still having trouble with keeping the aiming reticle on top of the surface and the trajectory line from drifting about. The feature still has some baking left to do.

Edited by Kobymaru

Share this post


Link to post
Share on other sites
1 hour ago, Argonicus said:

Any info when it will be updated to 1.3?

 

Share this post


Link to post
Share on other sites
1 hour ago, Argonicus said:

Any info when it will be updated to 1.3?

When it's ready and not before. I'm having a few coffees as I wait. :)

Share this post


Link to post
Share on other sites

Release should come soon-ish, if you can't wait then there's a pre-Release somewhere in the thread that should work with 1.3.

I'm on the fence about releasing the In-Flight Trajectory lines feature, because there is clearly some coordinate mess-up going on that I would like to work out. If I can clear that up, it'll be in the new version. If not, then I guess I'll release an "experimental build" for the early-birds that can help me debug the issues.

Share this post


Link to post
Share on other sites

New release is up!

Version 1.6.7 with KSP 1.3.0 compatibility, download from GitHub or SpaceDock.

Provided that the CKAN fairies have heard prayers and pleas, the update should propagate to CKAN soon™.

Full Changelog:

Quote

## v1.6.7 for KSP 1.3.0

- 2017-06-09

### Changes since the last release
  * This release is brought to you by fat-lobyte (aka Kobymaru on the Forums)
  * Update ToolbarWrapper for blizzy78's Toolbar. Fixes Issue #78.
  * Enable Trajectories window in flight scence.
  * Implement in-flight trajectory line display. Please test and give feedback!
  * Fix Parenthesis silliness in .version file to make it valid JSON again. Thanks, ggpeters!
  * Update Module Manager to v2.8.0. Thanks to PiezPiedPy!
  * Disable Click-through. Thanks to PiezPiedPy!

### Known Issues
  * Trajectory prediction is still fundamentally incorrect. See Issue #84.
  * In-Flight trajectory for non-atmospheric trajectorie s is jerky and jumpy.
  * In-Flight trajectory leaves targeting cross in the scene, even when disabled.

### For Developers
  * KSP directory path is not hardcoded anymore. You can set up an environment variable in your Operating System named KSPDIR that points to the KSP installation directory.
  * Revamped build scripts for Visual Studio. Thanks to PiezPiedPy!
  * Added simple profiler class. Thanks to PiezPiedPy!
 

 

Big Thanks to all contributors, especially @PiezPiedPy for their help!

Edited by Kobymaru

Share this post


Link to post
Share on other sites

Congrats! Glad to have both trajectory prediction tools together.

Share this post


Link to post
Share on other sites

@Kobymaru

The new version appears to maintain 1.2.2. compatibility, but I have a suggestion RE inflight trajectory. Could we get a toggle, so that trajectories can be on for the numeric (DNE data, thanks for adding it!) without the somewhat distracting line when it's not needed?

Share this post


Link to post
Share on other sites

I've encountered slight issues. Framerate drops a bit in orbit/flight and you always see a white cross, sideways, somewhere on the sky. It's small, but it's there even if I turn off the trajectory.

Nothing very concerning, but still an issue.

Share this post


Link to post
Share on other sites
14 hours ago, dlrk said:

@Kobymaru

The new version appears to maintain 1.2.2. compatibility,

Does it? I recall setting the lowest suppoted version to 1..2.9. Did I forget some config file?

14 hours ago, dlrk said:

but I have a suggestion RE inflight trajectory. Could we get a toggle, so that trajectories can be on for the numeric (DNE data, thanks for adding it!) without the somewhat distracting line when it's not needed?

Getting pretty crowded in the menu :wink: Could you report the issue on  GitHub?

5 hours ago, lajoswinkler said:

I've encountered slight issues. Framerate drops a bit in orbit/flight and you always see a white cross, sideways, somewhere on the sky. It's small, but it's there even if I turn off the trajectory.

Yes, the cross being always there is admittedly annoying. Could you also create a GitHub Issue here?

Share this post


Link to post
Share on other sites

I got strange bug. I have ship on orbit. I click to turn on trajectories and game crush.

I reproduce that bug twice. If need i can put output log and error log.

Share this post


Link to post
Share on other sites
Guest
This topic is now closed to further replies.