UO Second Age
http://forums.uosecondage.com/

Pet Commands: All <target>
http://forums.uosecondage.com/viewtopic.php?f=8&t=21190
Page 3 of 5

Author:  Kaivan [ Tue Aug 17, 2010 1:53 am ]
Post subject:  Re: Pet Commands: All <target>

As far as any information regarding all [target] commands suggests, the command worked like this during T2A:

A player controls 3 cats. Each cat is given a name; a, b, and c respectively. The player says all kill and targets a rat. The server processes the command and stores the values like this:

c kill
b kill
a kill <--- target cursor is up for this pet

The tamer then clicks on the rat and a begins to attack the rat.

Later on, the tamer decides to cast night sight on themselves. They cast the spell and receive the cursor. Instead of targeting for the spell being cast, the server does this:

c kill
b kill <--- server processes this stored command in place of the command being given

In this instance, the next queued command is executed and b attacks the tamer (in the instance of casting a spell, this may also leave a spell in limbo where it is ready to be targeted but is unable to be targeted resulting in the "You are already casting a spell." bug, but this is not necessarily certain).

Finally, the tamer decides to heal another pet using veterinary. When the tamer targets with the bandages, instead of beginning a veterinary attempt, this is done:

c kill <---- server processes the final stored command.

In this case, the final cat, c, will attack the pet that the tamer was attempting to heal (unless it is himself).

In this way, the server "stores" commands and will use them up one by one in place of the normally expected target command for each pet until it runs out of stored commands to execute.

Author:  Faust [ Tue Aug 17, 2010 9:38 am ]
Post subject:  Re: Pet Commands: All <target>

Kaivan, that sounds right and was my understanding of the bug after reading the discussion comments about it.


GuardianKnight wrote:
I read it the same way he did, that the all kill we were going on about actually worked but had a deadly bug so later past our era was disabled to fix a bug.


Based on my understanding it wasn't actually disabled but a notification was added that it didn't work properly for targeting commands.

Here is a portion from my first response in this thread that discusses it two days after the UOR publish went live.

Another off hand comment, you now get a message when you use the "all kill" command that it doesn't work. I sit here puzzled, it never had worked, so is
this an admission that it is broken or OSI taking credit for a bug as a feature that avoids an exploit?

Reply: I believe the comment is something about "The 'all' option cannot be used with a targeting command" (such as kill or attack) But it does make the closest pet perform the command...


The reply is from someone in the google groups responding to the poster. Based on this piece of information the fix only added a message, not physically disabled it completely. However, the problem with the stored procedures was apparently fixed later on though.

Author:  dren [ Wed Aug 18, 2010 12:18 am ]
Post subject:  Re: Pet Commands: All <target>

SighelmofWyrmgard wrote:
Derrick, you have given me the perfect opportunity to propose an accuracy convention that has hitherto been not considered (or, my bad, has been proposed and rejected before ...?).

I've been thinking about this quite a bit since joining UOSA, while observing "fixes" being applied, and all of the discussion that revolves around era accuracy.

The convention I wish to propose I will call, "thematic accuracy", or "spiritual accuracy": this is simply considering *whatever* in the context of how it was intended to be implemented as a finished, bug/exploit-free feature, in our target era; I do not propose a process of refinement, or tweaking, of any of the associated mechanics, merely an extirpation of identified flaws/bugs/exploits; that OSI either never implemented any fix at all, or indeed applied a fix but only after our target era, is irrelevant in this context.

UO was always an on-going project: numerous features were published in an incomplete and/or flawed/bugged/exploitable form; strictly insisting on mechanical-era-accuracy, regarding certain features that are obviously flawed, is simply 1billion% asinine.

"Thematic Accuracy", I imagine, is entirely the foundation for UOSA selecting Nov '99 as its "true" target date in the first place: this specific time was identified as the "cleanest" of all T2A publishes.

A partial list of what can be considered in the context of Thematic Accuracy:

    -post-era OSI patches that fix era-accurate flaws
    -post-era OSI patches that balance new content that was itself published in-era, or even pre-T2A
    -client/server/connection issues that are tied to era technological limitations

I don't want to open a can of worms here: many things can not be assessed in terms of Thematic Accuracy, simply because we don't have access to the designers' notes of our era; additionally, just because "something doesn't work right (the way I want it, waaaaaahhhhh!)", does not make it eligible for "Thematic Accuracy" consideration.

Some specific issues, and their disposition under Thematic Accuracy (as per my opinion); this list is exemplary, not comprehensive:

    -Above mentioned "all"-commands bugs: OUT; they are bugs
    -Runebooks: OUT; this is new content (but, why bother changing this?)
    -Server-Boundary emulation: OUT; the associated technological limitations do not apply at UOSA (sorry, all thieves and PKers who have been drooling over dreams of abusing the associated exploit ...)
    -Post-era Archery fixes: IN; this is balancing new content that was published in-era
    -Facets/Factions/Chivalry/Necromancy etc., etc.: OUT: new content, post-T2A
    -Poisoning checks against user's Poisoning skill: IN (at least for consideration); this patch was intended to redress a detected imbalance in poisoned-weapon usage
    -"wanderlust" NPC vendors: OUT; this was identified as bugged behaviour

Potentially, there is a lot more that could be analysed in this fashion; unfortunately there are a few issues that can never be resolved, because there is no OSI fix to refer to, nor any official discussion addressing the issue (I think analysing Stealthing Armor restrictions would fall into this category); I would also like to caution other members of the community that certain "fixes" that OSI implemented didn't fix anything, and we can not ask Derrick & co. to spend their time replicating these ...

In any event, I am neither proposing a scheme for "making UO better": I am hoping that this idea of Thematic Accuracy can be used by the UOSA team to best evaluate what needs to be replicated here, and how.

Thanks to all who actually read all this,

SS


Bump, and yes Runebooks are a new feature out of the T2A era completely; but they are convenient and it's possible they might have been implemented in the T2A era had OSI not been bought out by EA. Personally, I think these fit fine in the Thematic Accuracy; and since they're already being used it would be hard to take them away.

Author:  Derrick [ Wed Aug 18, 2010 8:28 am ]
Post subject:  Re: Pet Commands: All <target>

dren wrote:
Bump, and yes Runebooks are a new feature out of the T2A era completely; but they are convenient and it's possible they might have been implemented in the T2A era had OSI not been bought out by EA. Personally, I think these fit fine in the Thematic Accuracy; and since they're already being used it would be hard to take them away.

Runebooks were introduced within the T2A era. Rune Books Nov 23 1999 1:12PM CST Rune Books

Author:  Reena Dae [ Wed Aug 18, 2010 8:30 pm ]
Post subject:  Re: Pet Commands: All <target>

Kaivan wrote:
As far as any information regarding all [target] commands suggests, the command worked like this during T2A:

A player controls 3 cats. Each cat is given a name; a, b, and c respectively. The player says all kill and targets a rat. The server processes the command and stores the values like this:

c kill
b kill
a kill <--- target cursor is up for this pet

The tamer then clicks on the rat and a begins to attack the rat.

Later on, the tamer decides to cast night sight on themselves. They cast the spell and receive the cursor. Instead of targeting for the spell being cast, the server does this:

c kill
b kill <--- server processes this stored command in place of the command being given

In this instance, the next queued command is executed and b attacks the tamer (in the instance of casting a spell, this may also leave a spell in limbo where it is ready to be targeted but is unable to be targeted resulting in the "You are already casting a spell." bug, but this is not necessarily certain).

Finally, the tamer decides to heal another pet using veterinary. When the tamer targets with the bandages, instead of beginning a veterinary attempt, this is done:

c kill <---- server processes the final stored command.

In this case, the final cat, c, will attack the pet that the tamer was attempting to heal (unless it is himself).

In this way, the server "stores" commands and will use them up one by one in place of the normally expected target command for each pet until it runs out of stored commands to execute.


Yes, this is exactly how it worked.

Author:  Derrick [ Thu Aug 19, 2010 9:15 am ]
Post subject:  Re: Pet Commands: All <target>

Reena Dae wrote:
Yes, this is exactly how it worked.

Thanks much.

There are a few weird targeting bugs similar to this that were present on OSI; one actually involved bringing up a house target, logging off, and then logging back on and being able to use that target on the newly logged on different character. These are someone difficult to reproduce on UOSA because the targeting system is somewhat more secure (un spoofable).

I think it would be appropriate to have these bugs in the spirit of what we are trying to do here; but for the topic in question all evidence continues to indicate to me that we have done the right thing with "all <target>" pet commands. The introduction of the stored command bug would be an era accurate annoyance, and I am not chomping at the bit to implement this.

Author:  Mikel123 [ Thu Aug 19, 2010 11:24 am ]
Post subject:  Re: Pet Commands: All <target>

Derrick wrote:
I think it would be appropriate to have these bugs in the spirit of what we are trying to do here; but for the topic in question all evidence continues to indicate to me that we have done the right thing with "all <target>" pet commands. The introduction of the stored command bug would be an era accurate annoyance, and I am not chomping at the bit to implement this.


Fully agree with this. No one uses "all kill" anyways, which is at least the same behavior as T2A. If it takes you more than 5 minutes to reproduce something that no one is ever going to see or use, except maybe accidentally... it's probably not worth it.

Author:  Faust [ Thu Aug 19, 2010 3:20 pm ]
Post subject:  Re: Pet Commands: All <target>

I have to disagree Mikel and think it definitely needs to be implemented for other reasons besides era accuracy. This didn't just apply to the kill/attack command when prefixed with all, but instead the whole array of targeting commands with pets. The commands should be stored when they are issued in this manner. The mechanic would otherwise be very inaccurate and produce a system that never actually existed.

Author:  Derrick [ Thu Aug 19, 2010 3:32 pm ]
Post subject:  Re: Pet Commands: All <target>

While true though, my point was that it will be exausting to implement this, and while it's certainly on the list of things to do, in terms of cost/benefit it's wearing concrete shoes. Where cost is time and benefit is % of the overall accuracy picture.

Author:  GuardianKnight [ Thu Aug 19, 2010 3:41 pm ]
Post subject:  Re: Pet Commands: All <target>

Pets store commands in this setup?

So in theory, could i train my pet to always attack a certain red just by having it targeted to someone previously?

Author:  Faust [ Thu Aug 19, 2010 4:00 pm ]
Post subject:  Re: Pet Commands: All <target>

GuardianKnight wrote:
Pets store commands in this setup?

So in theory, could i train my pet to always attack a certain red just by having it targeted to someone previously?


No, not really. Try to think of it like this GK...

Say you have 3 dragons named a, b, and c in this scenario and issue an 'all kill' command with all three pets in view.

  • Player: 'all kill'
  • Pet A: Processses the command and produces a target.
  • Pet B: Processes the command and produces a target that wipes the previous target process by A that gets stored by the pet.
  • Pet C: Processes the command and produces a target that wipes the previous target process by B that gets stored by the pet.

Okay, using this scenario... there are no other pets left to process the 'all' command and the last target is left up that was triggered by the pet named C. This pet will be the only one that responds since the previous targets were essentially wiped and stored on the pet. However, if you were to produce a new target icon rather it would be from targeting a bandage, casting greater heal/ebolt, or anything for that matter that simply creates a target the next stored command will be processed by the closest pet that stored their last command when that target is used. For example, in this scenario above pet B would follow his 'kill' command when a target is issued and pet A would finally finish his command on the next one, etc...

This would obviously be a nasty bug when a mistake is made by the tamer and would have severe consequences in specific instances. A good example of this would be when using a bandage on one of your other pets and instead of applying the bandage the other pets would attack it, etc... same could be said for healing yourself or even a friendly that would flag you criminal.

Author:  GuardianKnight [ Thu Aug 19, 2010 4:05 pm ]
Post subject:  Re: Pet Commands: All <target>

Thanks for the explanation.

If it does get implemented, we can keep our underdog status here.

Author:  Faust [ Thu Aug 19, 2010 4:14 pm ]
Post subject:  Re: Pet Commands: All <target>

This actually makes a lot of sense now due to the fact that UO's code is infamous for bad garbage collection with their targets.

The horrible 'you must wait until your last spell finishes.' bug is a great example of this process. One way to reproduce this bug was by overriding your current spell target icon by producing a new target. For example, cast an energy bolt and hold the target icon. Use an item such as a dye tub or scissors via last object macro/hotkey and you got the bug. This would simply over lap the spell target and the system still thought you were holding the spell due to bad garbage collection.

The 'all' commands when utilized with targets is pretty much the same situation.

Author:  dren [ Fri Aug 20, 2010 5:17 pm ]
Post subject:  Re: Pet Commands: All <target>

Derrick wrote:
dren wrote:
Bump, and yes Runebooks are a new feature out of the T2A era completely; but they are convenient and it's possible they might have been implemented in the T2A era had OSI not been bought out by EA. Personally, I think these fit fine in the Thematic Accuracy; and since they're already being used it would be hard to take them away.

Runebooks were introduced within the T2A era. Rune Books Nov 23 1999 1:12PM CST Rune Books


Thanks, my mistake I don't remember them being introduced then but maybe I'm thinking of the point when they were added to vendors.

Author:  Mens Rea [ Sat Aug 21, 2010 3:29 pm ]
Post subject:  Re: Pet Commands: All <target>

nope it was more proximate than what you are suggesting fasut.

if you said all kill thne targeted some person some pet would attack.

but if you then said all follow me one would attack you.

not if you casted night sight - unless you said all follow me.

i'll take you on faust, 1 vs 1.

Page 3 of 5 All times are UTC - 5 hours [ DST ]
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group
http://www.phpbb.com/