Updates For Testing

News and Announcements
Joueur Moyen
Posts: 175
Joined: Sat May 22, 2010 9:16 pm

Re: Updates For Testing

Post by Joueur Moyen »

"You will no longer auto-defend on weapon swings or damage alone."

Initial test of this is only after the player being attacked tabs out of combat.

For Player A (gray) attacking Player B (blue), with A using a macing weapon and B using a bow:

1. Player A attacks Player B from a distance and does no damage.

2. Player B autodefends with bow.

3. Player B tabs out of combat.

4. Player A moves next to Player B and swings, doing damage.

5. Player B no longer autodefends, having tabbed out of combat.

6. Player A tabs out of combat and both just stand there.


A different test for Player A (gray) attacking Player B (blue), with A using a macing weapon and B using wrestling:

1. Player A attacks unarmed Player B from a distance and does no damage.

2. Player A tabs out of combat.

3. Player A runs next to Player B and Player B autodefends with wresting, but Player A just stands there, having tabbed out of combat. (In other words, Player A does no damage at any time to Player B.)

4. Player A loses criminal flag and is blue to itself, but still gray to Player B from aggressor flag.

5. Player B attacks Player A and does not go criminal, but causes Player A to autodefend. (Player B was original victim, Player A is now blue but still an aggressor to Player B.)

6. Player B tabs out of combat and stops attacking Player A.

7. Player A tabs out of combat and stops autodefending.

(Now both just stand there.)

User avatar
Loathed
Posts: 675
Joined: Sun May 04, 2008 10:56 pm
Contact:

Re: Updates For Testing

Post by Loathed »

Faust wrote:Refer to the '00 patch notes.

"Saying “stop” to your pet will clear any previous commands." - Server Publish Jan. 24th, 2000.

The way this worked before this patch was extracted from the original osi code. The 'stop' command was the same as the 'stay' command much like how 'attack' and 'kill' were the same.

incorrect as saying all stay would not removed guarded tag from ur head, where stop would.

and also- isn't that patch a lil past our deadline?

Joueur Moyen
Posts: 175
Joined: Sat May 22, 2010 9:16 pm

Re: Updates For Testing

Post by Joueur Moyen »

Is the following supposed to happen?

1. Player A attacks Player B from a distance (wrestling) and goes gray. (Outside of guards.)

2. Player A tabs out of combat.

3. Player B autodefends (wrestling) by running next to Player A, but A just stands there.

4. Player B keeps Player A's aggressor flag active by occasionally autodefending, until Player A is blue to itself and a third player.

5. Player A and Player B run inside a guard zone.

6. Player B attacks Player A. (Player A's aggressor flag is still active, and so Player B stays blue to itself and a third player.)

7. Player B runs next to Player A and causes Player A to autodefend.

8. Player A goes gray and guards can be called on A. (Even though the criminal flag has long worn off.)

User avatar
Faust
Posts: 6247
Joined: Mon Sep 22, 2008 7:01 pm

Re: Updates For Testing

Post by Faust »

The patch is past out deadline and that is the point to be made with it. The 'stop' command did not function that way until after our cutoff date in other words... The command was just a replica of the 'stay' command before it was implemented in January '00 after our cutoff date.

Telling a pet to stop when it was attacked would only make him stand still(but he would fight back if the attacker was next to him) just like the stay command. A person would have to invis the pet in order to remove the attacker's combatant.

The stop command in its current form is era accurate.

User avatar
Loathed
Posts: 675
Joined: Sun May 04, 2008 10:56 pm
Contact:

Re: Updates For Testing

Post by Loathed »

hmm, why would invising the pet remove it? isn't it the other way around?

User avatar
Faust
Posts: 6247
Joined: Mon Sep 22, 2008 7:01 pm

Re: Updates For Testing

Post by Faust »

When you invis your pet the combatant status is lost from the attacker since he can no longer acquire the pet. However, you have introduced a concept that wasn't previously examined I don't think... For example, say you have a horse and it's attacked by a player. You cast invisibility on the horse and the player loses his combatant status with the horse. I'm wondering what validation checks come into play for a horse in this situation. Does the validation only check to see if the horse has a combatant or does it check to see if the horse combatant and the horse's combatant is the horse? This is something that should be looked into in the original code.

Joueur Moyen
Posts: 175
Joined: Sat May 22, 2010 9:16 pm

Re: Updates For Testing

Post by Joueur Moyen »

Horses will wander around after the stop command, but they seem to stay put after the stay command.

User avatar
Derrick
Posts: 9004
Joined: Thu Dec 13, 2007 7:49 pm
Location: Cove
Contact:

Re: Updates For Testing

Post by Derrick »

Joueur Moyen wrote:Is the following supposed to happen?
...
8. Player A goes gray and guards can be called on A. (Even though the criminal flag has long worn off.)
Yes, because A took a swing at a blue player. This would not happen if A tabbed out, or B went out of range of A.
Image
"The text in this article or section may be incoherent or very hard to understand, and should be reworded if the intended meaning can be determined."

User avatar
Batlin
wobble llama
wobble llama
Posts: 512
Joined: Fri Dec 19, 2008 4:32 am
Location: Ultima Universe

Re: Updates For Testing

Post by Batlin »

The command was just a replica of the 'stay' command before it was implemented in January '00 after our cutoff date.
I'm not convinced this is true. OSI's pet script suggests that "stay" affects "following mode" only, whereas "stop" affects both following and guard mode.

There are too many unknowns at the moment about the NPC AI to give a clear answer. Someone with some free time at his hand could play with pets in the demo and write down what's happening.

Joueur Moyen
Posts: 175
Joined: Sat May 22, 2010 9:16 pm

Re: Updates For Testing

Post by Joueur Moyen »

Derrick wrote:
Joueur Moyen wrote:Is the following supposed to happen?
...
8. Player A goes gray and guards can be called on A. (Even though the criminal flag has long worn off.)
Yes, because A took a swing at a blue player. This would not happen if A tabbed out, or B went out of range of A.
One test I had A run away and hide, and then come back and still go gray after autodefending.

I'll try it with B going out of range, hiding, and coming back and re-attacking. (With A blue to third player, but still gray to B.)

I tried to break LOS/aggressor flag by taking the moongate with B, but it dumped me way down in the SE part of the lost lands, lol. The Papua transporter wasn't working, so I bought a horse and ran down to the Brit entrance, which allowed me to recall back to Skara. (Glad I had teleport regs on that char!)

But the moongates might be a little buggy on Test.

User avatar
Faust
Posts: 6247
Joined: Mon Sep 22, 2008 7:01 pm

Re: Updates For Testing

Post by Faust »

Batlin wrote:
The command was just a replica of the 'stay' command before it was implemented in January '00 after our cutoff date.
I'm not convinced this is true. OSI's pet script suggests that "stay" affects "following mode" only, whereas "stop" affects both following and guard mode.

There are too many unknowns at the moment about the NPC AI to give a clear answer. Someone with some free time at his hand could play with pets in the demo and write down what's happening.
Back when this was tested it seemed like it was reacting similar to the stay command. Both the stay and stop commands wouldn't get the pet to stop fighting. I didn't do extensive testing on it besides trying different ways to get a pet to stop fighting and just assumed it was a carbon copy of one command much like the other commands already had in place. I don't think there is any true way of extracting all the information from visual testing though. The exact functioning would probably have to come from the code in my opinion. :/

I'm still trying to get the demo to work on my Win7 machine in order to start testing stuff again.

Joueur Moyen
Posts: 175
Joined: Sat May 22, 2010 9:16 pm

Re: Updates For Testing

Post by Joueur Moyen »

Derrick wrote: Yes, because A took a swing at a blue player. This would not happen if A tabbed out, or B went out of range of A.
Going out of range doesn't seem to matter

This test started with Player A outside guards, Player B inside guards, and Player C watching:

1. A attacks B from a distance (wrestling) and does no damage, then tabs out and hides.

2. B tabs out and hides, then waits for A to come out of hiding. (A's criminal timer in Razor is at 100 sec.) B then retaliates while hidden from a distance (wrestling), and does no damage.

3. A tabs out of combat and B tabs out of combat, with neither having autodefended, and A runs into guards.

4. B rides away (From Skara moongate to south docks) and hides for 20 secs.

5. B rides back and re-attacks A from a distance, doing no damage. (A is blue to C but still gray to B.)

6. A tabs out. (And faces B, lol.)

7. B rides next to A and punches A, but A does not respond, and A stays blue to C.

8. B tabs out, rides back to docks, hides for 10 secs, rides across the boat teleporter and down to the first farm, hides, rides back to the moongate and re-attacks A from a distance. (A still gray to B, but blue to C.)

9. B runs by A, causing A to autodefend.

10. A goes gray to C. (And also gets "guards can be called..." message.)


Here's a simpler test:

1. A attacks B, tabs out and hides. (A is now gray.)

2. B tabs out and recalls to Brit from Skara, and hides.

3. A runs into guards, still gray to C.

4. B recalls back (with 100 sec on A's criminal timer) and re-attacks A.

5. A, now blue to C, does not tab out.

6. B runs next to A, causing A to autodefend and go gray to C.

It seems as long as B (initial victim) keeps A's aggressor flag active, A will always go gray again if forced to autodefend.

User avatar
Derrick
Posts: 9004
Joined: Thu Dec 13, 2007 7:49 pm
Location: Cove
Contact:

Re: Updates For Testing

Post by Derrick »

Joueur Moyen wrote: Here's a simpler test:

1. A attacks B, tabs out and hides. (A is now gray.)
2. B tabs out and recalls to Brit from Skara, and hides.
3. A runs into guards, still gray to C.
4. B recalls back (with 100 sec on A's criminal timer) and re-attacks A.
5. A, now blue to C, does not tab out.
6. B runs next to A, causing A to autodefend and go gray to C.

It seems as long as B (initial victim) keeps A's aggressor flag active, A will always go gray again if forced to autodefend.
Aye, this is correct though, if A takes a swing at a blue character it's a crime, in town or not. My comment earlier about going out of range was on losing the other player as a "combatant" (in which case you would not wswing at him if he simply walks next to you), but if he comes back and makes you take a swing at him you are committing a crime, it is still you who initialed the combat (until the fight has ceased completely for two minutes)
Image
"The text in this article or section may be incoherent or very hard to understand, and should be reworded if the intended meaning can be determined."

Joueur Moyen
Posts: 175
Joined: Sat May 22, 2010 9:16 pm

Re: Updates For Testing

Post by Joueur Moyen »

Thanks, I wasn't trying to suggest that it was incorrect, just testing it and letting you know my results.

I don't think you broke anything so far. :)

User avatar
Derrick
Posts: 9004
Joined: Thu Dec 13, 2007 7:49 pm
Location: Cove
Contact:

Re: Updates For Testing

Post by Derrick »

Joueur Moyen wrote:Thanks, I wasn't trying to suggest that it was incorrect, just testing it and letting you know my results.

I don't think you broke anything so far. :)
Your efforts are very very appreciated. It takes me some time to wrap my head around the various scenarios. Criminal, Guard whacking, murder count, stealing, perma-grey, agressor flagging. It gets relatively complex fast.

Everything that was on test up to yesterday is published this morning at patch 132. Thanks to everyone who's helped with this!
Image
"The text in this article or section may be incoherent or very hard to understand, and should be reworded if the intended meaning can be determined."

Post Reply