Playerbot

[Gitch]

Added same code to playerbotmgr.cpp for anyone who uses the POINT emote to send bots to attack..

heres the patch IN ADDITION to the one posted above: (so two diff patches, one for each of two files)

Patch changed is tested okay

diff --git a/src/game/playerbot/PlayerbotAI.cpp b/src/game/playerbot/PlayerbotAI.cpp
diff --git a/src/game/playerbot/PlayerbotMgr.cpp b/src/game/playerbot/PlayerbotMgr.cpp
index 3471e92..d87420f 100644
--- a/src/game/playerbot/PlayerbotMgr.cpp
+++ b/src/game/playerbot/PlayerbotMgr.cpp
@@ -355,7 +355,13 @@ void PlayerbotMgr::HandleMasterIncomingPacket(const WorldPacket& packet)
                                        for (PlayerBotMap::iterator itr = m_playerBots.begin(); itr != m_playerBots.end(); ++itr)
                                        {
                                                bot = itr->second;
-                                           if (!bot->IsFriendlyTo(thingToAttack) && bot->IsWithinLOSInMap(thingToAttack))
+                                           if (!bot->IsFriendlyTo(thingToAttack) && !bot->IsWithinLOSInMap(thingToAttack))
+                                           {
+                                                   bot->GetPlayerbotAI()->DoTeleport(*m_master);
+                                                   if (bot->IsWithinLOSInMap(thingToAttack))
+                                                           bot->GetPlayerbotAI()->GetCombatTarget(thingToAttack);
+                                           }
+                                           else if (!bot->IsFriendlyTo(thingToAttack) && bot->IsWithinLOSInMap(thingToAttack))
                                                        bot->GetPlayerbotAI()->GetCombatTarget(thingToAttack);
                                        }
                                        return;

[Gitch]

@Kennumen,

hey guy, Looking over the autobot code, and trying to wrap my head around what you were doing with it .

I have some ideas for independent bots. Not sure if you want anything added to autobot, but anything I

might come up with you're welcome to. I may just create a separate alternative branch as what I have in

mind may not fit the main goal of the main code.

Do you mind if I gather ideas from autobot?

By the way guys, come check out the new playerbot forums

(link in signature)

[blueboy]

... the test to see whether the bot can equip the item's armour class and not greater (i.e Paladin can wear mail; if the item is leather or cloth it will return false as it's not the best armour for the bot).

Yes sorry it will work with plate armour too, but if I had mentioned plate in my example, someone would have pointed out that this is only available at level 40 Before, the paladins having all armour classes would grab everything. If you were equipping a paladin manually, your more likely to choose mail over leather and cloth. Once available, plate would be the preferred choice

Hope that's clear

[Gitch]

Oh boy.. you're as wordy as I am...

sooo, with all these worms you let outta the can, I gotta address a couple

Okay, it's been a while and I'm a wordy sort so this could turn out to be a mammoth of a post. I'll try and keep it short.

blueboy wrote: ... the test to see whether the bot can equip the item's armour class and not greater (i.e Paladin can wear mail; if the item is leather or cloth it will return false as it's not the best armour for the bot).

Yes and no. Of course, at level 40 paladins and warriors can wear plate (and hunters and shamans can wear mail). But that is a complicated area. It's not as if at level 40 a warrior is suddenly clad in plate only. It's not as if a plate item is suddenly better than a mail one (and the issue is even more clouded for e.g. hunters who even later on often prefer a great rogue item (leather) over a -possibly lowerlevel- mail one). As a general rule, you want some plate items by 45 (ish) and to more or less be full-plate by level 50. More or less the same for melee shamans. Hunters, healer-paladins, spell/healer shamans are again more difficult. If they're in a group, they're never really meant to be hit, in which case it could be well worth it to exchange a bit of 'useless' armor for better stats. It's a cloudy issue and I haven't (and probably can't) explain it fully, but it's not always clear-cut (depending on class+function) that a lower armor class means the item is automatically 'unfit'. That said, it does enrage some people when e.g. hunters roll on leather, so if nothing else it's certainly a good option to have.

This is slowly creeping more towards stat decision than just level/armor class. Those were a good base to start from (and lower levels rely mostly on these).

Theres still a lot of tweaking to do but, one thing to remember, is that they are bots, and the autoequip function isn't going to take place of precision equipping.

I'd say, for alot of folks, they'd be fine just letting the bot equip whatever just to keep from having to do it themselves lol. However, for the hardcore player that

wants things to be just so, autoequip won't be a tool they'd turn on and forget about.

Still, yeah, it needs alot more work im sure, and atm its really mainly in need of testing, thinking, tweaking, and draw a line somewhere as to just how decisive it can be

and still be flexible for the players style/ideas/own way of doing things.

Gitch wrote: this is working out rather well at the moment.. Ive added a sub command for follow (follow dist) may change that but for now.. thats what it is.. You type it, and bot(s) will increment their follow distance by a set number, up to three times.. I'll explain, /p follow dist Bot whispers -"my follow distance is now set to medium" /p follow dist bot whispers -"my follow distance is now set to high" /p follow dist bot whispers -"my follow distance is now set to highest" /p follow dist bot whispers -"my follow distance is now set to Normal"

Why not a mixed command? /p follow dist [distance] where [distance] is (e.g.) a number between 5 and MAX_FOLLOW_DISTANCE (which I believe is a config setting and/or a declared constant in playerbot/playerbotmgr). OR [distance] is (e.g.) close (say 5), normal (default or 10), far (15), further (25), furthest (MAX_...). bot whispers "I'm now trailing you by # yards" unless it's exactly the value of one of the string variables (regardless of whether a # or string was the actual command), then (e.g.): /p follow dist 10 bot whispers "I'm now trailing you by 10 (normal) yards" Something like that. Just an idea.

Ahh this has changed alot from that post. follow -same as it was; follow far-increments distance +1.0 (with no current cap);follow near-decrements distance -1; follow reset- resets to config default;

the text on this can get kinda spammish... i guess.. I want to do alot , ALOT of work on the texts I've added to the code.

I havent put a cap on it yet because of testing ( as I said I'm looking into independant bots that will go about doing their own thing and also interesting to see how bots react ad different distances, which might be useful for formations during battlegrounds? just thoughts)

Gitch wrote: config file setting for "levels lower than bot" to tell bot how low level of a white item he/she should auto sell when told to. Default will be 10 levels lower than bot. (I like 5!)

I've been thinking about this for a while (how many of you have actually gone 1-80 at least once - just a question out of curiosity with no intended condescension), but shouldn't the white-item selling be dynamic? At level 1-6 you barely ever want to sell a white. Hell, you barely ever get a white. At levels 12-20 you're starting to get a fair amount of greens, and with quest rewards should be starting to equip quite a few greens(assuming, of course, the player does quests and wants to take the time to complete his bots' quests as well - easy for kill quests, time consuming for collection quests). Past level 20 you're not going to want many whites, excepting the shoulder position where grays are even acceptable until level 30 (or even higher if no replacement has been found yet). It's pretty much guaranteed by level 40 whites are useless - but I can't for the life of me remember if they even still drop by then. Excepted by this are of course tabards and shirts, but those probably aren't classified as armor anyway. Anyway, just thinking out loud as to whether one setting is enough. On the other hand, if it's not auto-equipped, is a white really worth the effort?

At the moment, my bots are level 18, and when they sell, they sell around 10 -13 items or so and approx half of that is from this code. Saves me alot of clicking ( and I have the level diff set to 4 for my own personal likes.) but I added the config setting so a server OP could change, or even disable it. I may instead opt for a user chosen setting, which ... now that I think about it would make more sense. if you're at a level that you're not getting alot of white items, this code isnt only white items, but any food/drink items that dont meet your level setting.

Future of this function: going to make it so that the player can set, from in game, what the bot will NOT sell, and see how that works.

Thanks for the feedback about autobot , and everything else too!

My idea for independant bots ( I believe that it can be done *mostly* from using already existing functions in playerbot, given the proper new section for ai, with only minor changes to the rest of the code to check if its an independant or regular bot)

The first project I took was: make bot find a nearby questgiver (easy enough, that function already exists) make the bot take the quest automatically (also in code already) make the bot recognize what the quest requires, creatures, items etc(already in code) Bot scans the map (temporarily using a MUCH larger radius heh) finds the creature, gets best path, goes to it, kills the creature,gets the quest item. When it has enough, returns to quest giver, does automatic turn in.

Most all of the functions needed for that is already in code. Having done this, getting the bot to continue doing it, from one quest giver to another, is alot easier. Then other tasks could be added, timeouts could be given, all to create variety in the bots behavior. bot could join groups etc, and after a random period of time, that bot will "log out". after another period of time, the bot logs back in.

yeah, maybe that all sounds crazy, but I warned yas earlier on that Im really really into scripted ai haha.

I promise I won't let this draw my attention from the more "needful" tasks of playerbot. I'm here to try to help!

Kennumen, get well!

I promise, I take no condscension from constructive criticism!

I admit, Im a slow programmer, and I have Not played through to 80 haha, I love to play around with crafting, questing etc, and

exploring, AND trying different races/classes, so i wind up starting over quite often!

I also love "creating" more than I do playing the actual game :lol:

So guess what I spend more time doing 8o

[Gitch]

New branch -chatbot .

Short explanation - branch will hold statements/responses for bots to randomly use. As it grows, it can be merged with main code in stages.

Bots will randomly say things, and others will respond based on a number of conditions such as class/sex/race and situational chat.

If you're interested in helping or offering up your own statements/responses to be included in the code,

Show your creativity and post your suggestions at the post HERE

Please keep it clean =)

[Gitch]

UPDate - Pushed to skillbot

Auto Start Quest Items - bots will automatically 'use' items that start quests when they get them.

(and of course accept the quest that it starts!)

If you don't want the quest, simply drop quest & resulting item if any. This is part of the AutoEquip function now.

If bot has already used the item, has the quest, AND loots another, They will remark that they don't need it, and loot no more.

(AutoEquip has to be on of course for this to work) 8o

Also added Uncommon Items to the 'sell all' command. Any Uncommon items that meet the low level requirement you set in the

config will be sold if you give the 'sell all' command prior to selling.

Take care!

[Gitch]

AutoEquip, and Adjustable Follow Distance are going to get plenty of attention next week probably, as I want to finish tweaking the itemstatcomparison and polish them both up for a tentative push to the master branch.

After the tweaks to autoequip and some added code to adjustable follow distance , which will make it somewhat automatic in the way bots change distance, they will be near a point that going

further might make matters worse instead of better :lol: Unless something else comes up!

Afterwards, I'm hoping to begin implementing AutoCraft, which will be built around autoequip, and its other functions.

bots will automatically craft certain things that go with their tradeskill, like leather from scraps, bolts of cloth from linen and wool etc.

And even automatically craft items for each other perhaps, we'll see.

anyway, just a note about whats up next.

for more details check out the Playerbot Forums.

[kennumen]

Discussion is clearly shifting towards the other playerbot forums so I'll keep it short.

... the test to see whether the bot can equip the item's armour class and not greater (i.e Paladin can wear mail; if the item is leather or cloth it will return false as it's not the best armour for the bot).

Yes sorry it will work with plate armour too, but if I had mentioned plate in my example, someone would have pointed out that this is only available at level 40 Before, the paladins having all armour classes would grab everything. If you were equipping a paladin manually, your more likely to choose mail over leather and cloth. Once available, plate would be the preferred choice

Hope that's clear

Fair enough. I mentioned elsewhere there is a transition period (41-45, at the least) and there are exceptions (a hunter may desire rogue equipment - without a rogue/feral in the group there's really no objections). For your example, I've seen paladin healers don cloth on occasion (although I believe it was WotLK that added decent leather/plate healing equipment - there wasn't any for quite a while). But I think we're more or less on the same understanding.

This is slowly creeping more towards stat decision than just level/armor class. Those were a good base to start from (and lower levels rely mostly on these).

Theres still a lot of tweaking to do but, one thing to remember, is that they are bots, and the autoequip function isn't going to take place of precision equipping.

I'd say, for alot of folks, they'd be fine just letting the bot equip whatever just to keep from having to do it themselves lol. However, for the hardcore player that

wants things to be just so, autoequip won't be a tool they'd turn on and forget about.

Still, yeah, it needs alot more work im sure, and atm its really mainly in need of testing, thinking, tweaking, and draw a line somewhere as to just how decisive it can be

and still be flexible for the players style/ideas/own way of doing things.

This is how so many people feel and I just can't understand it. Fine, some people enjoy tweaking and micro-management. I agree they want an option to set little modifications to their own taste. But why are people expecting the player to be not understanding of his own class, but understanding of the other 4? Why would I care whether it's better for a paladin healer to have mp5 or spirit? I just want the best chance to stay alive. Is it better to have strength or defense or agility for a tank? How does dodge compute against whatever else? I don't want to know. I want one someone else to figure it out, put it in playerbot, and make it so the evil thingies hit the tank and not me. If Joe Superman decides it's better for his level 80 tank to have a 2 hander, so be it.

As for ideal, I do realise it's quite complex, but given a set of parameters there is always at least one ideal solution, and several great ones. Not to mention, good, average, and acceptable

That was a bit long but it's something I feel passionate about but have seemingly been unable to convince anyone else.

Ahh this has changed alot from that post. follow -same as it was; follow far-increments distance +1.0 (with no current cap);follow near-decrements distance -1; follow reset- resets to config default;

My suggestion stands

At the moment, my bots are level 18, and when they sell, they sell around 10 -13 items or so and approx half of that is from this code. Saves me alot of clicking ( and I have the level diff set to 4 for my own personal likes.)

Perhaps it would be prudent for me to refrain from further comments on this matter until after testing.

Your independent bot sounds great, I've always planned to make it part of autobots. One more question on that though - where are you getting the 'characters' for your bots? Surely you're not doing automatic questing with alts?

[Gitch]

Some things are cut and dry, but not always. If I want to run to the store, I may take the car, or I may get in the truck, Or I might hop on the bike. They will all get me there, but not all the time. It might be raining, or muddy, or snowing, or just cold, but I still like my bike, and keep it even when It won't do the job.

There's really no perfect answer for every situation, in wow the same applies. If someone wrote the code, to cover such decisions, even just one person, no one else involved, the code would never get

finished. Choice after choice would pop up, begging to be reconsidered until the coder pulled his hair out haha. There are so many items, (and far too little time) to try to cover every situation.

True, that *most* of the time, you would not be faced with such hairsplitting decisions. I'm counting on the *most* of the time, BUT where it does require more factors to be decided, or compared, I've added a number of comparisons, and have continued to add more as I noticed an issue, or someone mentioned one.

The code can only make the decision that the coder tells it to make. If all the theorycrafters all had to put their theories into the code, without exception, my money says it wouldnt compile haha.

No offense to the theorists! Myself, I hate math! (don't ask me why I'm trying to code, or why I deal with engineering for a living!) I just want to have a function that will allow me to micromanage when I want, and when I don't want to, will handle some of the more menial tasks. I love a good RTS, but sometimes, I don't want to think so much.

I'll openly admit, I don't have the patience to produce the code necessary to make the perfect decision on items, if there is one. It's obvious what the most important factors are for each class *most* of the time, and the code even in its current state, is already checking for alot of them, and has been since probably the second push of autoequip (maybe 3rd!)

I can't help but feel I misunderstand your concerns, or perhaps you might misunderstand what the function currently does and how it does it? And with the current code in the works, there are even more class specific checks in place already, with more as I go. This might partially be to such a long conversation taking place over so many posts in a single thread as it has. I might not have noted the changes enough as I posted each push as I went. I should be more specific I know. I'm hoping the new forums will help address issues like that, but I still feel keeping an update here is good too.

I can't vouch for anyone else, but personally, its the micromanagement of each of my chars simultaneously , that drew me to playerbot. Even with autoequip, I still have control over what they equip If I disagree. However, its fun to be able to turn on some automatic functions and watch them choose on their own. (adds a bit of depth!) It's almost as fun as being beside myself! and In that I am unanimous.

In another game entirely, in a different time, I had a wizard. I refused to ever equip him with a robe. (and never did) out of protest for the ridiculous design and thought that went into them.. it was as though they said.. oh yeah, we forgot caster armor and just came up with something at the last minute... Now, Alot of people laughed about it, and mocked, and at the end of a fight I had ALOT less mana than other wizards.. but I still had mana. In fact, I was never out of mana (when playing conservatively as all casters should who rely on a tank to keep the mob off them!)

Now, perhaps thats extreme, but a good example. In the end, I think autoequip (if all goes well) will make good enough choices that no fight will turn out bad based on the equipment the bot is wearing. It still won't be perfect however.

However, My opinion aside (and the code isn't written on my opinion!) anyone is free to adjust, make better, add to, take away, whatever, with the code. I'll do the best I can.

Oh yeah, almost forgot! Yeah I'm doing the testing atm with alts. Its easy to manipulate them to run different tests.

[kennumen]

Some things are cut and dry, but not always. If I want to run to the store, I may take the car, or I may get in the truck, Or I might hop on the bike. They will all get me there, but not all the time. It might be raining, or muddy, or snowing, or just cold, but I still like my bike, and keep it even when It won't do the job.

Analogies are often dangerous, but hey let's run with it. Whether I take my car, truck or bike, there is still a shortest route to the store. There are several other routes which are also acceptable. Driving in the opposite direction will also get me there, 25000 miles later. ... It's probably best not to think of a tightly-knit city's road network, for this analogy. At least for my point of view

I can't vouch for anyone else, but personally, its the micromanagement of each of my chars simultaneously , that drew me to playerbot. Even with autoequip, I still have control over what they equip If I disagree. However, its fun to be able to turn on some automatic functions and watch them choose on their own. (adds a bit of depth!) It's almost as fun as being beside myself! and In that I am unanimous.

When I wrote that part of the response I had in mind high-end level 80 endgame. Tough 5man heroic instances where people are looking to squeeze out that last 2-3%. What I keep hearing is everyone wants to fine-tune every bot/class themselves. Personally, I just want to play my character myself. Suppose I want to go to a 10/25man raid. Sure, I may want to go as a corruption warlock [talent spec] when destruction is where the real raid DPS is at. At this point I really don't want to figure out what the [up to] 9 other classes should spec as. I don't want to figure out what might be their best gear. I just want to have a fair chance at having fun. Full party wipes, time and again, are -generally- not fun. I like research enough to do it for my own character - what's a decent talent spec for my playing style, what's some good gear, ... But I really don't want to do this extensive research for 4 (or even more) characters.

Of course, the fact that you do is part of what makes you a great developer for those features lol.

Having said that, some of that thought process is keeping in mind autobots. By definition they need to be good enough without any outside help (otherwise there are cheating issues).

Oh yeah, almost forgot! Yeah I'm doing the testing atm with alts. Its easy to manipulate them to run different tests.

I'm curious, do you plan on keeping it that way? Personally I would see altbots who do quests on their own as a cheat. I would more see this as a zone-filling feature for autobots. Either way, cool feature

[Gitch]

Gitch wrote:

Oh yeah, almost forgot! Yeah I'm doing the testing atm with alts. Its easy to manipulate them to run different tests.

I'm curious, do you plan on keeping it that way? Personally I would see altbots who do quests on their own as a cheat. I would more see this as a zone-filling feature for autobots. Either way, cool feature smile

No no, Just for testing! While the thought of hitting the auto follow and letting my bots do all the work sounds fun, it would definitely seem like cheating to me ] Its being done in a completely separate directory away from skillbot. At the moment thats all in , I guess you could say planning stages I'm trying to be careful and not get too many ideas rolling at the same time or I'll wind up with a bunch of ideas and nothing that works. However, talkin about them leads to better ideas for the current work.

I've done only a "little" coding to test out ideas and kind of .. feel it out a bit.. Zone filler is exactly what I have in mind. I'm basically testing code itself. Existing code. When I started thinking about what I would want them to do, It began to dawn on me that.. hey.. that code is already there... so then the main coding will be a wrapper to automate it! (thus all my "Auto" ideas! 8o ) Well.. more than that, but thats the gist of it.. No they'll definitely not be alts, but would be groupable etc, they would just .. somewhat randomly, level themselves, do trades perhaps.. but I will start it off simple to see how it goes.

If its compatible with what you have in mind for Autobot, great! I still haven't really delved into autobot to check it out, but I will when I work more on these ideas

If my post last night seemed a bit testy, It was a long day and I was pooped, and probably should've waited to reply this morning :lol:

(no testiness was actually intended!)

Take care

[Gitch]

One note, the biggest issue facing independant bots is the fact that someone.. will have to implement pathing for bots.. at the moment, they move in a straight line to their destination, which includes running through world objects and even under the world to get there. (just like pets will) I don't want to disturb the motionmaster code or any of that, so ... that part might get tricky. Might have to write a separate pathing function for bots.

[Gitch]

Updated - Pushed to skillbot

ItemStatComparison now compares not only the class/stat, but the stat value. If bot is a caster and item A has 5 int and item B has 6 (assuming no other stats exist on either one)

Bot will choose item B. (other stats on either item would similarly affect the outcome)

tidied up code (restructured the ItemStatComparison code to get ready for individual class/stat importance comparisons)

A fix in the automatic quest item use that led to the bots thinking they got another when in fact they didnt.. and they kept telling you about it..

Added Kreegoths Wand code for warlocks, priests, and mages

[kennumen]

If its compatible with what you have in mind for Autobot, great! I still haven't really delved into autobot to check it out, but I will when I work more on these ideas

At its core, the only thing autobots stands for is a bot that isn't based on an existing character. All else (whether the server or user asks it to login, what are it's main intended purpose(s), ...) is variable.

At least should be once I bloody get it working. Sessions, opcodes, etc... It's a bit over my head tbh, but what I do have so far I know is working, and on the right track.

<edit>

Added Kreegoths Wand code for warlocks, priests, and mages

oh, and no, no you didn't. Pretty sure, at least.</edit>

[Gitch]

it's in there

ICE_ARMOR               = ai->initSpell(ICE_ARMOR_1);
   ICE_BLOCK               = ai->initSpell(ICE_BLOCK_1);
   COLD_SNAP               = ai->initSpell(COLD_SNAP_1);

   // RANGED COMBAT
   SHOOT                              = ai->initSpell(SHOOT_2);

   RECENTLY_BANDAGED       = 11196; // first aid check

   // racial

This is just from the mageai.cpp but all the files include the code. works fine on my end.

Is it not working right? I've only tested with the current party I have.

[Gitch]

Update: Fix for Warlock Ranged Wands attack (this file missed getting Kreegoth's new code)

I have an issue with the warlock files opening from within the same project directory, Remembering that you have a Warlock, I double checked, and yes, I had forgotten to copy the edited warlockai.cpp over.

So it figures you found it :lol:

Those two filels refuse to open within their own project, and i've yet to determine why.

Thanks for letting it be known!!

ahh I overwrote it with the last push (prior to this update) what i missed was copying it to the MAIN "working" directory, and then turned around and overwrote the good one before I pushed...

Thanks for bringing that up before I wound up with a worse issue.

[kennumen]

Oh I see what went happened. You pushed it, then commented on adding it in the next commit, but that commit accidentally removed it from the warlock.

[Gitch]

Yeah, I just fixed this issue I hope, I removed the two Warlock files from the project, and then readded them to "game" and at least for now, I can open then.

Probably messed up something with a "cancel build". Anyway, thanks again for catchin that.

[kreegoth]

Hmm is this where I come in after some crazy work hours the last few days and say heh heh at all the confusion it seems ive caused LOL..

So any feedback on the wands? Preferred to have it working? Preferred to not have wands being used? any odd behavior noticed?

[Gitch]

Haha, nah, you actually caused me to find a bug in my solution with those warlock files.. it may have saved some more serious issues down the road.. seems I had corrupted my makefile or something.

All good now.

As far as how its working, I haven't had the chance to try it other than on my mage and priest bots, but it works fine on them.

Other than that I've been trying (feebly) to help blueboy chase down a bug in skill learn and work on itemcomparison while toying with an idea or two on autocraft. I'm hoping to get to do some better testing the next couple days!

**edit**

actually blueboy fixed the skill learn bug, the bug being currently chased is black with red spots and its called [skinning].

next push will hopefully contain the foundation for autocraft, though It may or may not be useable, I'll have something to build on in later pushes.

I have the outline code laid out for chatbot, once I tweak the structure a bit, I'll start adding convo's and see about a test run.

[kennumen]

I don't know if it's in new-ai or skillbot but I have a very annoying bug with auto-sell. When I get to a vendor they don't auto-sell anything, and don't even list their sellables. I have to open trade to get a list - the sell command works fine however, but I have to use it on everything. Would rebuild to locate the problem but on a 3day mini-lan atm.

[Gitch]

Kennumen, do you know when you last pulled from either new-ai or skillbot?

(so i kind of have an idea where to start tracking)

I'm checking it now to see if its working or not with the skillbot code. I havent merged new-ai into it in the past week, so not sure if theres an issue between them or not.

**

Edit**

its working here..

A thought comes to mind.. What vendors are you having problems with?

Blueboy fixed a bug last week where bots weren't recognizing some vendors. it could be related.

Just in case, I'll be pushing skillbot this evening with a few changes (nothing major) and I know the source is working fine

for auto sell.

[Gitch]

Update Pushed:

AutoCraft is working... actually much much better than I hoped for.. that worries me haha..

Sooo.. I'm going to tentatively push later tonight, but I'd recommend being careful about using it until its tested further.

At the moment, its not selective about what it creates.. It searches the bots items, and if it finds a tradeskill item, it will try

to make the first recipe it comes to that uses that item, and make as many of that recipe as it can..

I'll be working on how selective it is in the coming days..

See Here for details

[Unkle Nuke]

If leveling the bot's profession is the idea, you could always have the AI select which items to make based on skill level for that profession.

Gray - 0% chance an item will be made.

Green - 20% chance

Yellow - 40% chance

Orange - 80% chance

You could also add a command so a player may direct a bot to make a specific item, in case you do want to crank out a few lower-level items for sale at the Auction House or to help a friend. Maybe something like botname craft [item name].

[Gitch]

You could also add a command so a player may direct a bot to make a specific item, in case you do want to crank out a few lower-level items for sale at the Auction House or to help a friend. Maybe something like botname craft [item name].

Hey UnkleNuke !

Blueboy has already taken care of the commands for crafting specific item recipes, in fact my code is based on his work

You can type /t botname craft [recipe] and the bot will craft it, or craft [recipe] all, and the bot will craft all of that recipe that it can.

I appreciate the feedback, and I like the idea of using skill level as a basis of whether or not the bot will create that item or something better!

Unless its needed for a higher level recipe (such as an item that creates a better one) theres really no reason for a bot to auto make it if its Gray.

Unless of course asked for one.

Thanks again!

By the way, the Wiki has been updated a little, check out the features link for a break down of the different commands etc.

its still a work in progress.