Kallor
Level: Master Delver
Rank Points: 277
Registered: 04-02-2007
IP: Logged

Okay maybe an oversight rather than a bug but I thought this best fits under this topic anyway.

Alex Diener realized in his youtube let's plays that pushing moves don't activate saving spots. Playing around the editor I verified that this is also true for tokens, which is a slightly more serious matter.

I find this a little counterintuitive, but I understand it's probably too late to change this. But architects watch out, forcing the player to use a token might not work because of this.

____________________________
Kallor

blorx1
Level: Smitemaster
Rank Points: 920
Registered: 07-18-2009
IP: Logged

This is because the player and other pushed things are treated as objects while being pushed.

This doesn't seem to prevent triggering fuses however.

____________________________
If you need to think outside of the box, then you didn't build a good enough box.

disoriented
Level: Smitemaster

Rank Points: 2401
Registered: 08-07-2007
IP: Logged

Just to be clear, we're talking about a player being pushed with a pushing weapon, right? Not when the player is pushing a keg or mirror (?)

____________________________
34th Skywatcher

Best to PM me, since I might miss your message on CaravelNet chat.

Jutt
Level: Smitemaster
Rank Points: 863
Registered: 06-29-2007
IP: Logged

That seems indeed what Kallor meant. I am getting some interesting results testing this. What I found so far:

- The player pushing a mimic/decoy/etc onto a token does not activate it.
- A mimic or other entity pushing the player onto a token does activate it.

- The player pushing a mimic onto a checkpoint does not activate it.
- The player rotating their sword making a mimic push them onto a checkpoint does not activate it either,
- However: the player taking a step and then being pushed onto the token on the same turn does activate the checkpoint.
- Also note that the player stepping onto a checkpoint and then being pushed off by the mimic on the same turn does not activate it. It does make sense only to activate a checkpoint when you stand on it at the end of the turn, because that is when the state of the room is actually saved.

I'm not sure what was intended, but I agree with Kallor that the most intuitive behaviour would be to always activate tokens. For checkpoints it would be ok to only activate them at the end of the turn, but still activate no matter how the player landed on them.


____________________________
Holds: An Architects Audition, Artful Architecture, Salamander, Elusive Exhibitions, Leftover Levels, Six Times Six
Collaborative: Way Forward, Advanced Concepts 2
Styles/Mods: Basalt, Sandstone, Garden, Clock using game elements

Dischorran
Level: Smitemaster

Rank Points: 3414
Registered: 09-10-2005
IP: Logged

I wonder if this is related to this issue that was giving me so many scripting headaches.

____________________________

The spitemaster
Level: Smiter
Rank Points: 354
Registered: 06-09-2005
IP: Logged

This is a fix for my bug. It involved being pushed off time tokens on the turn that you stepped on them. The game would instantly crash if you tried to return to them. These might be unforeseen side effects.

____________________________
Last night upon a stair
I met a man that wasn't there
He wasn't there again today
I wish that man would stay away

bomber50
Level: Smitemaster
Rank Points: 872
Registered: 09-18-2006
IP: Logged

File: Pushing Checkpoint Inconsistency.hold (1.1 KB)
Downloaded 46 times.
License: Public Domain

The fix for 5.0.2 is currently bugged. See the attached hold.

adS
Level: Master Delver

Rank Points: 214
Registered: 05-02-2010
IP: Logged

bomber50 wrote:
The fix for 5.0.2 is currently bugged. See the attached hold.

What is CCW?

best wishes,
adS

____________________________
pauca sed matura

Dragon Fogel
Level: Smitemaster
Rank Points: 2467
Registered: 06-21-2014
IP: Logged

Counter-clockwise. The way you turn with the Q key.

disoriented
Level: Smitemaster

Rank Points: 2401
Registered: 08-07-2007
IP: Logged

Most likely known as "anti-clockwise" in adS' part of the Eighth.

____________________________
34th Skywatcher

Best to PM me, since I might miss your message on CaravelNet chat.

adS
Level: Master Delver

Rank Points: 214
Registered: 05-02-2010
IP: Logged

disoriented wrote:
Most likely known as "anti-clockwise" in adS' part of the Eighth.

Well we mathematicians call it positive.

We even call 180 degrees pi

adS

____________________________
pauca sed matura

skell
Level: Legendary Smitemaster

Rank Points: 3764
Registered: 12-28-2004
IP: Logged

bomber50 wrote:
The fix for 5.0.2 is currently bugged. See the attached hold.

Unless I am missing something the both cases are intended behavior:

Case #1, CCW -> Restart:
- The game starts, Turn 0, nothing happens.
- CCW move, Turn 1, Beethro is pushed around, at some point the checkpoint (24, 18) is triggered, Beethro ends up on (16, 18).
- Restart, since the previous checkpoint was triggered on Turn 1 and we are right now on Turn 1, restart back to the beginning. Player is taken back to (25, 18), back to Turn 0.

Case #2, CCW -> Wait -> Restart:
- The game starts, Turn 0, nothing happens.
- CCW move, Turn 1, Beethro is pushed around, at some point the checkpoint (24, 18) is triggered, Beethro ends up on (16, 18).
- Wait, Turn 2, nothing happens.
- Restart, since it's turn 2 right now we are taken back to Turn 1 when we hit the last checkpoint, player is moved to the end moment of Turn 1, (16,18).

I think the issue is that it looks weird for the checkpoint to be triggered at (25, 18) yet the player ends up on (16,18) but it's all intended - it's not important WHERE is the checkpoint you've triggered. What is important is the end state of the turn the checkpoint that was triggered because, internally, restarting to a checkpoint replays all of the moves from the beginning of the room up until the turn when the checkpoint was triggered.

Well... Unless the room is demonstrating something that has already been fixed in 5.0.2.1004 and I just can't see it, so please correct me if I am wrong!

____________________________
My website

bomber50
Level: Smitemaster
Rank Points: 872
Registered: 09-18-2006
IP: Logged

Yes, it just looks weird.

skell
Level: Legendary Smitemaster

Rank Points: 3764
Registered: 12-28-2004
IP: Logged

Alright, then that is intended! Can you confirm this thread is fixed then?

____________________________
My website

bomber50
Level: Smitemaster
Rank Points: 872
Registered: 09-18-2006
IP: Logged

Confirmed in 5.0.2.1004.

Dragon Fogel
Level: Smitemaster
Rank Points: 2467
Registered: 06-21-2014
IP: Logged

Bump to request clarification. So it seems this was settled for checkpoints, but I don't see any comment that the current behavior is intended for tokens.

To reiterate things said before: that behavior is that a pushed mimic or decoy does not activate tokens, but a pushed player does. This has significant consequences. It would be more intuitive (and less prone to breaking rooms) if a pushed double always activated tokens. I can't think of a room in a published hold that depends on this behavior.

But if this is intended behavior, oh well. I just want to make sure it is.

skell
Level: Legendary Smitemaster

Rank Points: 3764
Registered: 12-28-2004
IP: Logged

I agree it makes sense that pushes should be consistent. I don't foresee this breaking "too many" things but who knows? Anyhow, fixed PR

____________________________
My website

mauvebutterfly
Level: Smitemaster

Rank Points: 771
Registered: 05-03-2015
IP: Logged

Just to be clear, does pushing a double onto a token now activate the token? This will definitely break a bunch of rooms in Entry Point, and probably in other holds that have been published in the 4 years since Fogel's last comment.

If this is definitely what's happening then that's fine. I'd just like extra confirmation that everything didn't break when you did this and that I need to be changing some of my rooms.

____________________________
106th Skywatcher

skell
Level: Legendary Smitemaster

Rank Points: 3764
Registered: 12-28-2004
IP: Logged

Yes, that's the idea. I understand that it might break some stuff and ultimately we will see the scope of damage done by this when the test spider runs and rechecks all the demos.

If it breaks some rooms in holds that are still in development I don't very much mind, they are in development for a reason .
If it breaks a lot of rooms then... we'll have to make a call whether to keep the change or revert it. So don't be too hasty with changing rooms in developed holds yet.

Usual disclaimer that I am open to discuss the change and can be override by Mike.

____________________________
My website

---

[Last edited by skell at 10-09-2020 03:00 PM]

Nuntar
Level: Smitemaster

Rank Points: 4756
Registered: 02-20-2007
IP: Logged

Please don't change this. It's been established for a long time now; changing it will break a lot of rooms; it's not obviously wrong; and it has a lot of puzzle potential that would be lost.

____________________________
50th Skywatcher

---

[Last edited by Nuntar at 10-09-2020 03:35 PM]

mauvebutterfly
Level: Smitemaster

Rank Points: 771
Registered: 05-03-2015
IP: Logged

Just did a test with lemmings in the editor to test how the pushing rules apply to npc characters that activate tokens.

Lemmings normally activate tokens when they step on to them, but when pushed onto a token they do not activate them.

I'm not sure if this is something that could be changed with scripting, or if it's a hardcoded engine thing, but it kind of makes sense that npc characters would work the same as doubles too if we're changing this.

____________________________
106th Skywatcher

mauvebutterfly
Level: Smitemaster

Rank Points: 771
Registered: 05-03-2015
IP: Logged

Talking in chat, and skell said that the change would be for both doubles and npcs.

I personally think that doing both or neither is better than just changing one of them, so I was happy to hear that.

I guess I lean towards not changing this behaviour at all at this point since it feels more like "behaviour" than "bug" to me, and I'm already used to how things currently work. I'm not too bothered either way though.

____________________________
106th Skywatcher

navithmastero
Level: Smitemaster

Rank Points: 621
Registered: 01-03-2012
IP: Logged

Fixing this 'bug' I'd wager would break a sizeable number of rooms.

It also, as far as I can tell, would literally make a room I made impossible. See the secretted tags below. I only remember this room in particular as I made it, I'm sure there are a bunch of other rooms out there which will be rendered impossible by this change.


Changing this would make a precedent that we are happy to just go about changing established behaviour which has been used by people and has potential to break lots of rooms for no reason that I can tell, other than that it feels slightly strange. This is not a QoL issue or any sort of obvious breaking of otherwise established DROD rules. The purpose of bugfixing is to fix bugs, not change behaviour that we don't agree with.

Edit: this is in no way intended to be a dig at the amazing work Skell is doing trying to fix bugs and introduce new features to the game. I feel strongly about this issue however, hence my tone in the above message.

____________________________
Member of the Snake Appreciation Society

One of your local HAs.

My stuff:

---

[Last edited by navithmastero at 10-09-2020 04:47 PM]

Dragon Fogel
Level: Smitemaster
Rank Points: 2467
Registered: 06-21-2014
IP: Logged

Noting this because it came up in chat: If we want consistency, it might break fewer things to instead make it so pushing the player doesn't activate tokens.

The current behavior in that regard is mainly useful for enforcement mechanisms that use tokens to prevent the player from stepping in a particular place but don't block off monsters, but 1) there are other ways to do the same thing (including crudely with scripting) and 2) this particular one isn't very common (at least not in rooms that also have pushing weapons).

skell
Level: Legendary Smitemaster

Rank Points: 3764
Registered: 12-28-2004
IP: Logged

navithmastero wrote:
Changing this would make a precedent that we are happy to just go about changing established behaviour which has been used by people and has potential to break lots of rooms for no reason that I can tell, other than that it feels slightly strange.

I disagree very strongly. There are good reasons to claim this behavior is inconsistent (a. pushed player triggers tokens; b. moving doubles trigger tokens; c. besides tokens, pushing triggers every other room interaction I can think of at the moment). There have also been precedents with changing established behavior (tar baby spawn order is the biggest one I recall) and if you look at all the things we're fixing for 5.1.1 there are a couple of behavior changes that, I'd say, are similar on the scale of inconsistency. I usually try to judge behaviors by their discoverability, (which in this case I'd say is mixed), consistency (which in this can I'd say is bad), puzzle potential (I think both ways have merit) and intuitiveness (which in this case I'd say is low) - that's why I pursued this bugfix in the first place. I digress.

The fact it can break a lot of published rooms is an actual problem that I think is much more worth pursuing. Or that it will introduce a lot of unintended solutions.

Ultimately though, it's not up to me to make a decision- I feel like a lot of people have been trying to convince me to change my mind in the chat, and I want it's best to keep the meat of the discussion to this thread here as to not rob the people who are not in the chat from being heard.

And while I could force my opinion by just deleting the commits (chances are no one would have the time to reinvent the changes) I'd rather leave it up to Mike now. If it was up to me I'd keep things as they are given the weight of the arguments so far.

But the code is written, PR done, I'll leave it up to Mike to make the final verdict!

Btw, I'd appreciate feedback on other fixes that modify behavior, there are many changes making its way to 5.1.1

____________________________
My website

mrimer
Level: Legendary Smitemaster

Rank Points: 5134
Registered: 02-04-2003
IP: Logged

I appreciate the discussion points here, and it appears there has been an extended conversation on this topic as well in Chat that I've missed.

My two cents, fwiw:

It's not by design that npc/double pushed entities don't activate tokens when they otherwise would, either by placement or independent movement. I consider this an oversight in the implementation, not merely a strange behavior. Please correct me if I'm wrong, but my understanding is that there haven't been any definitive statements made to the contrary at any time.

I personally wouldn't agree that rooms where npcs/doubles are pushed onto tokens, but don't activate them, is an established behavior. Though development hasn't been too active for a few years, this topic has been an open bug report. I do value and respect architects' creations and it's a shame when rooms are broken by logic changes. However, there is a long-established precedent that rooms built on buggy behavior are at risk of being broken when the bugs are fixed.

I understand these need to be considered on a per-case basis. If there are rooms in Entry Point that rely on the current logic, we do need to consider it is possible this logic will change.

____________________________
Gandalf? Yes... That's what they used to call me.
Gandalf the Grey. That was my name.
I am Gandalf the White.
And I come back to you now at the turn of the tide.

Dragon Fogel
Level: Smitemaster
Rank Points: 2467
Registered: 06-21-2014
IP: Logged

There was a discussion in chat a few years back that seemed to settle things in favor of maintaining the current behavior - skell was involved in that discussion but I suppose he's forgotten about it. I only vaguely remember it myself, so I can't really offer any more detail than that.

I should note that one major effect of the current behavior is that it is impossible to change a decoy's weapon after the decoy has been placed. This also applies to scripted entities that activate tokens and do not move, such as spinning mimics.

This is used as part of the enforcement in The Island of Doom. Possibly other rooms involving decoys and weapon tokens, though I'm not sure offhand.

Oversight or not, this would not be a small change. It's also not obviously wrong - pushed entities other than the player get stunned, so it could be explained as "being stunned by a push prevents you from activating a token". (However, pushed NPCs that cannot be stunned don't activate tokens either, so that seems reasonable to change; I doubt that would have any impact on published rooms.)

Given the potential impact at this point and the fact that it's at least debatable as to what the correct behavior would be, I think it would be best not to change the current behavior.

skell
Level: Legendary Smitemaster

Rank Points: 3764
Registered: 12-28-2004
IP: Logged

Someone in the chat mentioned that changing this would affect hundreds of published rooms. Could someone spend the time to try to gauge the real scope of this change?

____________________________
My website

Doom
Level: Smitemaster

Rank Points: 3303
Registered: 07-05-2004
IP: Logged

skell wrote:
Someone in the chat mentioned that changing this would affect hundreds of published rooms. Could someone spend the time to try to gauge the real scope of this change?

As an example roomsearch gives over 100 results for rooms with a staff token and either a mimic or a decoy potion. Add other tokens and pre-placed NPCs and hundreds of potentially affected rooms sounds about right. How many would actually break is hard to say. Looking at random room images it doesn't seem like most of them would be affected much, but in this many rooms it's almost guaranteed that there will be some breaks.

Dragon Fogel wrote:
This is used as part of the enforcement in The Island of Doom. Possibly other rooms involving decoys and weapon tokens, though I'm not sure offhand.

For the record I'll have no hard feelings if this is changed. This room is already broken in other ways due to the so-called "succor" trick as mentioned here. Not to mention I don't have that strong of an emotional attachment to it.

mrimer
Level: Legendary Smitemaster

Rank Points: 5134
Registered: 02-04-2003
IP: Logged

Thanks for the additional context. Super helpful!

What do we feel next steps are? Merge the change, evaluate room breakage, and back out the change if the impact is subjectively egregious?

____________________________
Gandalf? Yes... That's what they used to call me.
Gandalf the Grey. That was my name.
I am Gandalf the White.
And I come back to you now at the turn of the tide.

---

[Last edited by mrimer at 10-13-2020 11:39 PM]