Kerbal Construction Time/StageRecovery Dev Thread

[JeffreyCor]

No on mission control.

- - - Updated - - -

And of course now that I try to reproduce it, it doesn't happen.

I'd have that oh so many times. I feel your pain

[Sof]

Don't know if this has been mentioned, but trying to select the crew with this mod is awkward.

It's even more important now if you're playing career that you have the right Kerbals in the command pod, but you have to randomize the crew several times in order to get the right astronaut. There doesn't seem to be any way to manually select the right man for the mission. It defaults every time to Jeb for me even if I selected another kerbal in the VAB/SPH beforehand.

When it comes to launching a vessel, you lose all options to manually select the pilot and the flag/logo for a particular mission.

[magico13]

I would suggest reading the manual prior to posting, though the album on the first post of the KCT page includes pictures of the crew select GUI You can (and always have been) able to select individual kerbals by removing the kerbal from the seat and selecting a new one (I'm on mobile, but it's either "select" or "add", I don't recall).

As for selecting a specific flag, you should be able to do that at the time the vessel is added to the build list. You may be able to edit the vessel and then choose a new flag, then save the edits to use the new flag. I don't choose individual flags for missions, so I've never tested it.

[eddiew]

Hey Magico,

Just wondering if you know any reason why KCT simulations might not have the same 'on rails' physics as a regular run?

Reason for query; I have a 2-part Duna ship, which leaves its (TAC) life support and fuel supply in orbit before descent. After ascent, it picks it up to come home. My simulation is to move the docked lander/orbiter as a unit to Duna orbit, split, and go from there. I've tried this scenario two ways with different results:

1 - using a KCT simulation; separation and descent fine and ascent out of atmosphere are fine, but gets weird after that. The orbiter never seems to be where the map view says it is, and all manoeuvres veer off in the last few m/s to point directly away from Duna. Trying to rendezvous seems to be like hitting a moving target; an intercept of 1.2km in half an orbit won't be an intercept by the time I get there. If I somehow fluke getting to within a few km, then burning to zero on the target-retrograde marker just results in the marker moving and the speed difference increasing. Feels a lot like I'm chasing something that's not in the same frame of reference, or is at a different point in game-time. (I always check whether I'm looking at orbital, surface, or target navball modes.)

2 - using HyperEdit to cheat the combined ship to orbit; perfectly normal behaviour all the way.

This may not be your mod's fault at all, but I'm getting desperate enough to ask, and it only seems to happen during simulations ^^; I have no problems doing simulated rendezvous (rendezvii?) around Kerbin.

Thanks!

[magico13]

@eddiew I don't know of anything off the top of my head that would cause that sort of behavior. Could have something to do with the transport into orbit (which is derived from the HyperEdit code), but if it works around Kerbin then it should work around Duna. You can try sending me a log as it might be an error that's being logged there, but I really don't know.

[eddiew]

Yeah, figured it might be something like that, Magico - thanks for replying though

Tbh, I might just live with the oddity. Tried to work out which mod is at fault and all I know is the answer is "some of 'em". Since it doesn't seem to affect 'live' gameplay, I could probably just ignore it.

*edit* Oh, I did actually have some issues with the initial orbital transport; it's too circular! The orbit created has a PE and AP that flicker uncertainly. This makes it impossible to leave something on that orbit and return later, because the game isn't sure where it is... Obviously my current node-marker-running-away bug is kind of making this irrelevant, but adding a small amount of eccentricity to that initial orbit would probably help

Edited January 12, 2015 by eddiew

[eddiew]

Hello again... right, I'm really sorry, but I'm pretty sure there is something odd about KCT simulations around some planets. I stripped the game down to basically stock (MJ included because I'm lazy with docking), and here's what happens with a simulation:

For comparison, I simulated on the Kerbin launchpad, then Hyperedited out to Duna. None of the above problems happened.

I've uploaded a log including the simulation run here although I'm not sure how much help it will be... I still don't know for sure that this is a KCT bug, there's an equal chance it's HyperEdit (since you said you use their code), or something about KSP itself that reacts weirdly to what KCT does before entering the simulation.

The good news (for me at least ) is that KCT is not breaking the 'real' world of KSP, and all oddities are confined to its simulations. My Duna mission shall launch tonight!

Apart from this little hiccup, a fine mod and definitely one I don't want to play KSP without! Although I may have to cheat on the sims now and then when I need to check out a Duna scenario

Edit: is it possible that the initial orbit, being 'too circular' for the game and making the PE and AP get confused and wobble around, is upsetting it from that moment onwards?

Edited January 12, 2015 by eddiew

[magico13]

I'll take a look at the logs, but I'm not sure if I'll get a chance to do in-game testing soon. Regarding the orbits being too circular, I find that timewarping can help solidify their position, alternatively a short burst of RCS or the main engine.

The node moving away is something I see in the stock game all the time as (presumably) a result of floating point inaccuracies, but probably not to the extent you're describing. Hyperedit recently updated (prior to that it hadn't been since 0.21) but the code I'm using is all based on the old code, so I'm not sure if they changed any of the things I referenced.

Edit: Just looked at the log. Not a whole lot there. No errors that I can see. Could you enable the Debug Logging in the KCT settings please, that way future logs contain more KCT specific output. Secondly, I don't know if it will help, but if when you go to set up the simulation you check "Show advanced options" and then check the "Delay move to orbit" option, does the same issue occur? That will cause a 3 second delay where the ship sits on the pad with the physics loaded, then is transported to orbit, rather than happening immediately when loaded. There's a chance that explains why HyperEdit doesn't have the issue.

Edited January 12, 2015 by magico13

[Agathorn]

I

The node moving away is something I see in the stock game all the time as (presumably) a result of floating point inaccuracies, but probably not to the extent you're describing. Hyperedit recently updated (prior to that it hadn't been since 0.21) but the code I'm using is all based on the old code, so I'm not sure if they changed any of the things I referenced.

KSP uses a floating origin to help with precision. Is there any chance that the code you are using to move to another planet's orbit isn't taking the origin with it somehow?

[magico13]

KSP uses a floating origin to help with precision. Is there any chance that the code you are using to move to another planet's orbit isn't taking the origin with it somehow?

I'm considering that. I'll have to look into it more, but the origin should move every 6km anyway. Though it might wait until the ship exceeds 6km from the previous origin, which this method might somehow prevent it from seeing (thus keeping the origin on the launch pad). Switching to another vessel from the map view, then switching back, should clear up that issue though.

[eddiew]

Yeah, I've been tapping the engine just to break those wavering markers, and I'm quite used to nodes that waver a little in the last 0.2m/s or so, but these are something else; if you manage to aim at one, it'll run away until it's pointing radially outwards from the planet, and simply can't be chased down. All sorts of weird stuff happens if you can get in visual range of your target too - it just isn't moving the way the navball says it is.

Ok, a log performing the above with VERBOSE_DEBUG_LOG turned on is here. I ran two sims with logging, the first without and the second with a delay on moving to orbit - both of them suffered the same thing with manoeuvre nodes running away while the ship is on-rails. I imagine the target movement is as crazy as described, but since they can't get near the target... xD

And don't worry if you can't look into it for a while; there's a workaround (launch to Kerbin and HyperEdit) and it's not affecting the game outside of KCT's own simulations, so no damage is being done. I'm just grateful that you created this mod at all, because it does add a lot to KSP. If a bit of my time gets spent helping debug/improve KCT, then that's ok

[magico13]

I don't know what VERBOSE_DEBUG_LOG actually does, I was talking about KCT's debug messages setting (in the KCT settings menu) (this menu, the Enable Debugging setting), but thank you anyway I don't think it's going to mention anything that will help anyway.

The stock issues I was mentioning are with the maneuver node marker moving up to 90 degrees away from where it said it should be while timewarping. Happens all the time in the stock game if your maneuver is a while away (the longer the time, the more it moves). Your issue sounds very much worse though.

Looking through the HyperEdit source, the section of code I referenced didn't actually change at all :/

[eddiew]

Pants, sorry, my mistake xD It'll have to wait until tomorrow for that I'm afraid, but I'll try to get round to it then ^^;

And mysteries abound it seems... I have seen nodes drift, particularly if switching off the ship and back, but... not like this does. I get the feeling the stock one is the ship slowly rotating, but MechJeb even has trouble locking onto these nodes to start timewarp because it can be sliding so rapidly :S

[magico13]

Don't worry about getting another log. I don't think there's anything that KCT will output that will tell me anything different, it's just nice to be able to see all the stuff it outputs if I'm going to be looking at future logs. It let's me know precisely which step it's at when KCT things are happening.

I might be able to do some stuff with Krakensbane (the thing that moves the origin) that'll help, but I don't really know how or where to do it. What happens if you start a simulation in orbit around Duna, then use HyperEdit to adjust it (staying within the system or changing SOI and then changing back, there might be different results for each). I imagine jumping out of the system and back would clear it up, but it's possible the initial jump made by KCT somehow throws everything out of whack.

[eddiew]

Interesting question - I shall try a Duna start with HyperEdit to a few other places when I get chance. Will let you know the results

[Flef]

I think I have an issue with KCT but I'm unsure.

I was testing the mod for the first time yesterday evening ('cause I find the mod is really what is missing to KSP career mode ).

I'm using the entire Umbra catalog, TAC-life support, Scansat, RPM, KSPRC, active texture management (ofc), Mod oriented tech tree, final frontier, FAR, Kronal Vessel viewer, contract window+, dangIt, Deadly reentry and the toolbar.

I'm using KSP in 32bytes

What happened?

I started a new career

I 've fulfilled two contract (launch a new vessel/reach 5000m altitude)

I didn't get the rewards (no money no sience added, probably no rep)

When I remove KCT the contract system is working properly.

Are the rewards taking time to be granted or was it a bug? Is it a particular setting I missed?

If it is a bug, what are the suggestions to make it disappear?

Edited January 14, 2015 by Flef

[magico13]

Are the rewards taking time to be granted or was it a bug? Is it a particular setting I missed?

If it is a bug, what are the suggestions to make it disappear?

Just to be clear, you did wait for the vessel to be complete and launched it by clicking the "Launch" button (after rolling it out to the pad, the "Rollout" button) in the KCT window, NOT BY CLICKING THE LAUNCHPAD? Someone previously had an issue because they simply weren't actually launching the vessel.

Assuming that's not the issue, I've seen this reported prior and it appears to be related (usually) to another mod having an issue, which then causes the save file to not finish loading. If you can get me the output_log.txt file I can take a look and see if anything is being reported there.

[Flef]

Just to be clear, you did wait for the vessel to be complete and launched it by clicking the "Launch" button (after rolling it out to the pad, the "Rollout" button) in the KCT window, NOT BY CLICKING THE LAUNCHPAD? Someone previously had an issue because they simply weren't actually launching the vessel.

...Was the issue, I clicked the pad...

Many thanks for your answer. You saved my KSP.

[magico13]

...Was the issue, I clicked the pad...

Many thanks for your answer. You saved my KSP.

No worries That might be the way to launch vessels in the future, but for now you have to go through KCT's windows. Glad the problem was an easy fix

[Mars Mullo]

SPH space plane refueling

Forgive me if I have missed a previouse post , I have searched this thread on refueling and SPH but...

I have been hitting alot of surveillance missions latley and I have an awesome little plane that does it. But I have not put alot of build points into the SPH and I have not spent 2 points to reset the points I have.

However I was wondering why if I land the plane and taxi it up to the hanger and all, why I can't just refuel it for the next mission?

I pressed recover vehicle and the aircraft is now gone... so I have to spend some 40 odd days to build a new one. (sure warping is no issue really)

The landed aircraft was in good nick... is there any game play method I'm missing?

I read the guide .. couldn't find any mention of SPH for quick turn around of aircraft of SPH builds.

Any help please?

- - - Updated - - -

OK I have been playing around with it a bit more.

Am I right that I can just click on the runway and load up the craft with kerbals I select.

And if it it shows the simulation greeting options but has no simulation clock countdown then I'm actually playing in game with the craft that previously landed?

[magico13]

If you see the simulation window at start, it's a simulation. There's a "secret feature" (read as "bug") where clicking on the launchpad or runway and launching that way will start an infinite time, cost free simulation.

How long did it take for you to build your plane originally? Was it 40 days? After recovering, if you go to build a new one and you use parts from the inventory (it's a plane, so all the parts will be in the inventory) so the build time will be 4 days instead, which is a reasonable amount of time.

To recover things straight back into the inventory is what you're more asking for. Unfortunately, converting an existing vessel back into a valid craft file (which is what KCT uses) isn't a trivial task. I did work on it prior to 0.90 and had a very buggy code that worked, that was subsequently broken by 0.90. There's been recent discussion in this thread that presents solutions I can use for making recovering straight to the inventory possible, but it will likely still have issues, especially with modded parts.

The alternative solution is to not recover the vessel at all. Simply keep a refueling tanker around and park it on the side of the runway!

[Mars Mullo]

A refueling tanker... hmmm I like that. Except that the KAS port will be the one part that will break my plane.. I'm at 30 part limit and I just swapped my KER unit for a solar panel because the plane runs down charge when landed for EVA.

Such decisions...

Ok found the bug twice now.. well at least I learned to fly half way around Kerbin and how to do it to use less fuel.. that is high and faster.

I could perhaps replace the two parachutes with one big one.

Yes the build time is some 40 - 50 days and after recovery was also about the same.

So I must not be using part inventory some how... will look into this closer the management on SPH builds is tricky.. but not to say this isn't fun.

It would be nice to have this info in the guide document though.. because I do look there.

Will the refueling tanker work for food and water too.. as I am using TAC.?

[magico13]

A refueling tanker would work for anything you fill it up with, so if it has life support supplies then you can transfer them. Also, KER doesn't need a part with the 1.0 series (though maybe with 0.90 you need the part if you don't have a scientist and/or engineer).

If you're not seeing the parts when you go into the editor and open up the part inventory (currently there's a bug that fuel tanks and engines won't appear), then send me your save file and a log file.

By the way, here's the relevant part of the Getting Started Guide that covers the part inventory and reuse mechanics

There is also a button “Part Inventoryâ€Â. Here you can see parts that have been recovered from

recovered vessels or boosters that had enough parachutes. Parts in the part inventory don’t take as much time to

build (about 1/10 of normal), just the time to fix them and attach them (probably with spit and buckloads of

duct-tape).

This means spaceplanes just take some hours to refuel, while rockets take days to build! Also, the

number of times you have used specific parts changes build times. If your engineers strapped some batteries a

hundred times, they will have no problem doing it again, while if it is a new part, they will take more time

reading the instructions manual! (After 4 launches a part takes about half the time to build, 16 = ¼, etc…)

[westamastaflash]

To recover things straight back into the inventory is what you're more asking for. Unfortunately, converting an existing vessel back into a valid craft file (which is what KCT uses) isn't a trivial task. I did work on it prior to 0.90 and had a very buggy code that worked, that was subsequently broken by 0.90. There's been recent discussion in this thread that presents solutions I can use for making recovering straight to the inventory possible, but it will likely still have issues, especially with modded parts.

For rockets at least, is there any way to maybe just cache the craft file and put it back in inventory if you haven't launched yet when you recover?

[magico13]

For rockets at least, is there any way to maybe just cache the craft file and put it back in inventory if you haven't launched yet when you recover?

Same issue as before, converting a vessel into a craft file. They've got radically different structures. However, you can revert to the editor and it will be in storage (since you're reverting back time)

Edit: I see what you're saying. That would be possible to do actually, but Claw's discoveries in the thread I linked to previously mean I might have a solution that's valid for all situations. We'll see.