Fix joint RID not being passed to _set in PhysicalBone

[Demindiro]

The joint RID was never passed to the _set() function of the joint_data, which meant the joints on the PhysicalServer's side wouldn't be updated.

Attached is a reproduction project (just the platformer demo but with ragdoll physics). Press tab to toggle between stiff and pudding mode.

I think it would be best to remove the default RID() value for the RID j parameter. It's probably better to be explicit to prevent similar bugs in the future.

@Favkis could you check if this patch fixes your issue please?

platformer.zip
platformer.zip (works just well enough for Godot 4)

[Chaosus]

Changes should be made for the master branch first - that's a general rule of our contributing pipeline. Could you please create a 4.0 version of this PR?

[Demindiro]

Changes should be made for the master branch first - that's a general rule of our contributing pipeline. Could you please create a 4.0 version of this PR?

I will once @Favkis confirms whether this change fixes his issue or not.

[Favkis]

@Demindiro
It works now! Thank you so much!

[Favkis]

It works. Now .set works for 6DOF and other types!

2020-12-20.15-45-04.online-video-cutter.com.mp4

As you can see on video, (upper torso is 6DOF), first run character watches up because torso is set to 89 min 90 max. Second time I comment the code and character falls relaxed.
Screenshot with values -125 and -126, works great, upper torso is bent.

[Chaosus]

You've redirect it to master? Never do that again - it's not a common practice + this will confuse other contributors which maintain the bug tracker, you should open a new PR that targets it.

[Demindiro]

You've redirect it to master? Never do that again - it's not a common practice + this will confuse other contributors which maintain the bug tracker, you should open a new PR that targets it.

Okay, I'll keep it in mind for next time.

[madmiraal]

Well done. As you suggest:

I think it would be best to remove the default RID() value for the RID j parameter. It's probably better to be explicit to prevent similar bugs in the future.

If you make that change too, I think it will be good to go.

[madmiraal]

I've added the cherrypick label, because, although the change is not cherrypickable, the change is easy enough to backport and it was originally created as a 3.2 PR.

[madmiraal]

@Demindiro Thanks for the update. Please squash your commits into a single commit.

[akien-mga]

Thanks!

[Favkis]

@Demindiro
Looks like that fixed only .set but not doing same through PhysicsServer:

[Demindiro]

@Favkis Can you check if print(rid.get_id()) returns 0? I'm fairly certain what the issue is but I'd like to confirm.

[Favkis]

@Demindiro

[akien-mga]

Cherry-picked for 3.4.

Looks like that fixed only .set but not doing same through PhysicsServer:

Was there a follow-up for this? If not, it would be good to open a dedicated issue so that it's not forgotten.