Possible Bug: Pressing Dig while climbing drops item in addition to letting go, with jump and run control

[fritzw]

Description says it all. Should there be a break at the end of this line?

LegacyClonk/src/C4Object.cpp

Line 3599 in a08f927

case COM_Dig: ObjectComLetGo(this, (Action.Dir == DIR_Left) ? +1 : -1);

I noticed this because my COM_Jump implementation copied that line and also dropped items when letting go. It seems unhelpful to have a key that does "let go" and "drop" at once, because you can collide with the dropped object, or fall into the explosion of the flint you just dropped. If you wanted to do both, just press both keys at once.

Interestingly, the DFA_Scale case in C4Object::DirectCom (line 3421) does not mention COM_Dig at all. It's only in C4Object::AutoStopDirectCom.

[maxmitti]

Yes, it looks like a bug (and it also annoyed me very much when trying JnR).
However, there are players who like it as a feature for letting go and dropping the object very soon …

[fritzw]

Well for my gamepad patch it does not matter much, because COM_Jump is a separate button, so you can just press A (Jump) to let go without dropping items or press B (Dig) to let go with dropping items.

For keyboard control I don't really care much. So if the bug has grown into a requested feature, I'd be fine with closing this issue.

However, players who want to do this could still press "Throw then immediately Dig" to achieve the same effect. It's not as if fast button combos are unheard of in Clonk. Just saying. In the end it's your call, I don't care either way.

[maxmitti]

However, players who want to do this could still press "Throw then immediately Dig" to achieve the same effect. It's not as if fast button combos are unheard of in Clonk. Just saying. In the end it's your call, I don't care either way.

If I remember correctly, the horizontal drop position is different when using Dig.

I once had an idea of creating an improved JnR scheme which would be the new default, but opt-in for existing players. The idea included fixing this. I don’t recall the other changes. Perhaps I can still find the discussion somehow.

[Danghorx]

Results of the poll from years ago. Noone liked the dropping of the items while scaling:

[fritzw]

I don't quite understand the first point in the survey. Neither the JnR controls nor the classic controls drop items on Dig or Down in the DFA_HANGLE case. Or was that different in the past and got fixed?

Quote @maxmitti
I once had an idea of creating an improved JnR scheme which would be the new default, but opt-in for existing players. The idea included fixing this. I don’t recall the other changes. Perhaps I can still find the discussion somehow.

I also considered implementing my Xinput gamepad controls, with an explicit jump button, as player->ControlStyle == 2, but then decided against it, because then I would have to change all those if (player->ControlStyle) throughout the code. In the end I decided against it, because there is only one place where it would have made a difference (in ObjectComUp), and all other places just work the same as before if COM_Jump is never sent.

On the other hand, I might still do it, because it would allow a bit more freedom to change the controls to be more gamepad/analog-stick optimized without impacting keyboard controls. But I don't think it would be worth the effort for just this one break;.

[Danghorx]

You drop items after falling down when you hangled, because it still has the JnR input for Down active for half a second. This sometimes leads to you dropping your weapon (like a spear or gun) or item on yourself that you would have wanted to throw or shoot. With classic this doesn't happen because you need a second down press for dropping an item.

[fritzw]

Ah yes, that makes sense. Luckily I didn't encounter that behavior yet in my tests. But I see how it would be seriously annoying in a fight, or when dropping a flint on a slope. Fixing that probably requires more than a break; though.

I'll definitely fix that in my gamepad patch. If you think it's a good idea, I can create a PR for the main branch with just those changes.

[maxmitti]

Thanks to @Danghorx ’s poll results, I think it is reasonable to fix it unconditionally.
If you want, you can create that PR, otherwise I can do it too, since it is a simple change. :)

[maxmitti]

Regarding dropping of weapons only, this would need to be fixed via Clonk and individual object scripts.
I wonder if people would have a problem if it’s fixed in general.
Especially, because the exactly same thing could still be achieved by holding down [Down] until [Throw] has been pressed.

[Danghorx]

If it is easily implemented, you could implement both JnR "Bug-fixes" and let it be tested by the players. And if it is accepted and doesn't break anything, it can stay this way. Otherwise you can revert the changes.
Funni proposed more DefCore settings but as the discussion below shows, Fulgen doesn't want it. It is more a problem of the engine than of the packs.
Whole discussion in discord:
JnR Bug Controls.txt

[Danghorx]

Is there a reason in the code why pressing dig while scaling or hanging lets the clonk go of the wall or ceiling? This happens only with JnR.

[fritzw]

It is there since the first commit to this repository, so I guess someone must have thought that the dig key is underused when first inventing the JnR mode.

[fritzw]

I've found another weird behavior connected to this: When hangling in a very low corridor and pressing Dig to let go, the clonk would start digging immediately after landing. But if the fall takes longer than C4DoubleClick, then C4Player::ExecuteControl will clear LastCom and the Clonk will not start to dig after landing...

This is clearly a bug, so I fixed it in the same commit.

[maxmitti]

I once had an idea of creating an improved JnR scheme which would be the new default, but opt-in for existing players. The idea included fixing this. I don’t recall the other changes. Perhaps I can still find the discussion somehow.

I couldn’t find the discussion, but Danghor’s and your insights reminded me of a few things.
My personal opinions were and still are to change (if there was no consideration about players used to the current JnR controls) behavior during Climbing and Hangling:

• No engine behavior for [Dig] during climbing and hangling. This means, neither let go, nor drop inventory contents.
• Pressing [Down] to let go doesn’t set the flag which is responsible for dropping things.
• I would even consider to require holding [Down], [Left] or [Right] to drop things mid-flight.

However, especially the last point is strongly biased due to my experiences. I started playing Clonk with Clonk Planet, which only had classic controls, so I always have been very used to pressing [Down] for stopping.
Whenever I would use JnR controls, I would still press [Down] subconsciously many times in order to stop. So I often dropped stuff unintentionally.
But anyway, I think having this kind of “memory” for having pressed [Down] recently for alternative actions is rather unintuitive, especially when holding down keys is already the norm for controls. Together with it being defined as “has been pressed in the recent X frames” it makes a very unpredictable behavior during fluctuating FPS.

I also considered implementing my Xinput gamepad controls, with an explicit jump button, as player->ControlStyle == 2, but then decided against it, because then I would have to change all those if (player->ControlStyle) throughout the code.

For this reason, I was considering to (internally) use a separate flag, or change ControlStyle to a bitfield.
This way, everything would default to the already established JnR behavior, except when explicitly checked otherwise.

For the sake of having a distinction between [Up] and [Jump] a similar idea could apply: Set a new flag when a dedicated [Jump] binding shall be used and only add the check at the point where you want the change.