List of Stock Bugs that Should Finally be Fixed for 1.1

[Dunbaratu]

This is a new problem in 1.0 and above. It did not happen in 0.9: Doesn't drop out of timewarp soon enough when hitting atmosphere anymore, leading to catastrophes some of the time.

I'm on a return trajectory from the Mun to Kerbin. The orbit is an elongated ellipse with apoapsis up near the Mun, and periapsis is showing as 28km in the Kerbin atmosphere. I timewarp to get to the point where it hits atmosphere. But instead of dropping to 1x timewarp somewhere between 65km and 70km like it used to, it jumps well past the first point of hitting atmosphere, and drops me out of timewarp well deep into Kerbin atmosphere already. With this bug, it drops you right into that thick air, and unpacks the vessel at a speed still going well over 3000 m/s, without the decceleration you would have had had it been physics simulating the time spent passing through the thinner upper part of the atmosphere first before it got this deep in. It also drops you in without having had the wind rotate you shield-first, and without having had the chance to decouple the service module, as would have happened had you approached that point under physics simulation. The result is often an explosion before you have the chance to hurry up and decouple everything upon regaining control.

It used to drop you out of time warp pretty reliably right at the edge of atmo. Now I have to drop it down to 5x or so, manually well above the atmosphere, and gingerly approach the atmosphere slowly at only about 5x until it hits, so that it won't blow way past the start of the atmosphere like that.

It's annoying because even though the above is a workaround, it's a workaround that's tedious because dropping out of timewarp well ahead of time and approaching the atmo gingerly means extra waiting time.

Edited May 15, 2015 by Steven Mading

[Sonny_Jim]

KSP always had problems when at high warp near atmospheres, although usually it skips past them rather than dropping you deep in the atmosphere. KAC is invaluable for making sure you don't zoom past periapsis.

[Warpedone]

I was disappointed to see that the orbit decay bug has returned - sometimes when warping for a long period the orbit decays as if the engines had been engaged retrograde, resulting in either a catastrophic spacecraft/scenery interface or hilariously ending up on a 180 degree inclined orbit if left too long. This has just wrecked my first 1.0.2 Duna shot. Grrrr.

[stevehead]

Fix the embarrassingly sloppy SAS system that points the ship at prograde, retrograde, normal, etc. My ships often wildly overshoot the point, then wander back to overshoot again. The ships have plenty of control authority, since I can move the ship into position without overshooting, so this just looks like bad control logic.

This is a good example of a feature that a mod does extremely well, but Squad's implementation is sub-par. The Smart A.S.S. of MechJeb works very well, and is pretty much the only reason I ever install MechJeb (I'll use the auto-maneuver executor occasionally while I go grab a sandwich). From what I understand the problem has something to do with Squad's SAS using static values for the PID values, while MechJeb dynamically changes these PID values based on the spacecraft (forget who mentioned this). I'm not asking for Squad to implement everything like the target orientation and surface mode vectors (although I find those extremely useful), I just don't want to have to wait for the stock SAS go back and forth trying to get my spacecraft aligned to the vector I have chosen, when it would have been less of a hassle for me to have done it manually.

Another example of an implementation Squad did that a mod does better is Enhanced Navball. The ghosting feature works very well, and I like the anti-maneuver node (it has its occasional use). Of course, this is not a bug, so I'll end that here before I get too off topic.

Edit:

Actually, the stock SAS isn't really a bug itself. It's more of an annoyance due to poor implementation.

Edited May 15, 2015 by stevehead

[arkie87]

This is a good example of a feature that a mod does extremely well, but Squad's implementation is sub-par. The Smart A.S.S. of MechJeb works very well, and is pretty much the only reason I ever install MechJeb (I'll use the auto-maneuver executor occasionally while I go grab a sandwich). From what I understand the problem has something to do with Squad's SAS using static values for the PID values, while MechJeb dynamically changes these PID values based on the spacecraft (forget who mentioned this). I'm not asking for Squad to implement everything like the target orientation and surface mode vectors (although I find those extremely useful), I just don't want to have to wait for the stock SAS go back and forth trying to get my spacecraft aligned to the vector I have chosen, when it would have been less of a hassle for me to have done it manually.

Another example of an implementation Squad did that a mod does better is Enhanced Navball. The ghosting feature works very well, and I like the anti-maneuver node (it has its occasional use). Of course, this is not a bug, so I'll end that here before I get too off topic.

Edit:

Actually, the stock SAS isn't really a bug itself. It's more of an annoyance due to poor implementation.

My Two Cents:

If you click prograde and the ship cannot point itself prograde, that is a bug (unless it's somehow your fault, like you adjusted your trim poorly or have off centered thrust)

[stevehead]

My Two Cents:

If you click prograde and the ship cannot point itself prograde, that is a bug (unless it's somehow your fault, like you adjusted your trim poorly or have off centered thrust)

I haven't had the issue of it not pointing correctly. If that's something people are experiencing, then yes, it is definitely a bug. I was referring to how it goes about aligning your spacecraft to the vector, which technically works as intended based on the code, but it's far from ideal.

Edited May 15, 2015 by stevehead

[Renegrade]

That's a very good list in the OP. I believe I've personally experienced each one (or at most, all but one) of those bugs, and it would be nice if they got fixed.

[arkie87]

I haven't had the issue of it not pointing correctly. If that's something people are experiencing, then yes, it is definitely a bug. I was referring to how it goes about aligning your spacecraft to the vector, which technically works as intended based on the code, but it's far from ideal.

i've experienced that sometimes. In space, ship will refuse to point directly at vector and comes to steady state like 15 degrees off. If i manually move it to center, it will get flung off center to the other side...

[Dunbaratu]

As for the stock SAS problem. There is a new mod that just showed up on kerbalstuff that lets you manually fiddle with the PID tuning parameters that stock SAS is using, to see if you can get better behavior out of them:

https://kerbalstuff.com/mod/765/PID%20Tuner

Of course you need different settings for different styles of ship, and that's why it's good to have an auto-tuner for PIDs, but in this case even without an auto-tuner at least this gives you some control, or at the very least can operate as a diagnostic tool to work out why it's acting like it is.

[BurnPrograde]

EVA: (3) Jetpack controls simultaneously operate ship controls, making it impossible to reenter vessel. *Pulls hair out*

[arkie87]

EVA: (3) Jetpack controls simultaneously operate ship controls, making it impossible to reenter vessel. *Pulls hair out*

not sure what you mean here? How can you be controlling EVA and ship at the same time?

[THX1138]

What about being unable to either hide or unhide a certain "flight" type (debris, pods, stations, flags, etc) from the map screen until you go back to the tracking station.

That's not exactly a bug. It's more of a suggestion. Try suggesting it if you want it added

I'm pretty sure it's a bug because if you Alt + Tab out and back in again, it's fixed and you can hide/unhide things like you're supposed to be able to.

Edited May 17, 2015 by THX1138

[nholzric]

I'm pretty sure it's a bug because if you Alt + Tab out and back in again, it's fixed and you can hide/unhide things like you're supposed to be able to.

Someone suggested you can only change the selection if the nav ball is up. This seems to work for me but since it was intermittent before I was paying attention to the nav ball I'm not 100% sure. Even if this fixes it, it is nonobvious enough to need "fixing" imo

[Windspren]

The one where moving a maneuver node around deletes it is because if you move it behind your ship and past it it makes the game think the maneuver node already happened.

[RyanRising]

There is a vital question for me, it has been a silly matter since years and I feel like the moment to finally face it has come:

STRUCTURAL ADAPTERS

Bi-adapter, Tri-adapter, Quad-adapter...the first time you see those parts your eyes are shimmering of joy, but when you try to use them you recognize they do not adapt anything. We all well know that the structural nodes are not joining, just one of them actually does; so you are forced to use weak docking ports to have a sort of structural joining between parts.

Come on Squad. Seriously. Please. Pretty please.

How much hard the task can be. Have the structural adapters be structural for real.

So looped connection nodes? I'd love it, but it sounds really hard to do.