help unsticking player

[kx9488]

i have until recently been running a very old version of rombot that worked slower than this version but had little problems with being stuck.

the problem i am facing now, with the latest version is the unstick timer and counter.
"unsticking player from.... at position (x/y). Trial # from maximum 10 trials"

first of all, the bot keeps track of every single unstick trial so if it "unsticks" itself and moves on to another waypoint successfully, and later gets stuck again, the timer starts at trial #2 instead of at #1.

second of all... the bot initiates unstick trials in areas it isn't stuck. this usually happens after running to a waypoint with mobs. it kills the mobs, and instead of moving on the camera sways left to right a few times and then initiates unstick trials. most of the time it takes one trial and moves on... but in conjunction with the first issue my bot deosn't stay running more than 15 minutes.

i did change the camera degree from 15 to 25 like in this post http://www.solarstrike.net/phpBB3/viewt ... =21&t=2185 but that didn't help, i still get a lot of unnecessary unstick trials (there are literally no physical obstacles in the way... it isn't trying to run through walls, trees, over hills etc. the next waypoint is right behind my character or it is often standing almost on top of where it is supposed to be)

anyone else experience this or know why this is happeneing?

an obvious thing to do would be to change the maximum unstick trials to a larger number but when i really do get stuck ill be humping an obstical for like 20 minutes non stop and get banned.

i'll keep playing with camera angle degree and waypoint placement like i have been, before i give up and go back to the slower old version.

-thank you in advance

***edit***

i've noticed that the trial counter does go back down to #1 only if enough time has passed... not if it gets unstuck and reaches a new waypoint, this is still creating a problem.

also, if the unstick trial initiates and a mob spawns i enter an attack sequence and it saves me from the unstick trial almost every time.
when it unsticks itself from the unstick trial (running around in a circle and jumping up and down) it usually saves itself when it runs over the waypoint it was looking for just by sheer luck... it never needs to relieve itself from an actual obstacle... for some reason the problem occurs finding a waypoint almost always inches away.

***edit***

i managed to keep my waypoints the same and use angle 35 for that line of code int he player.lua ... i don't know if it actually did anything but my bot is working longer than it has been... going on approx 25 minutes.... if it gets shut off prematurely i sappose i could increase the unstick timer to 15-20 so it has more chances to not mess up for a while to get the unstick counter back down to trial #1....

still would like info on how to change the unstick counter settings to reset upon finding a new waypoint

[lisa]

I'd have to double check but from memory the unstick counter resets if you travel more then 3 waypoints.

I have noticed this issue of getting stuck for no apparent reason, as yet I haven't been able to pinpoint the actual cause. I don't use quick turn as you mentioned in your other topic. I notice it seems to happen in certain places though, bot will run smooth and then this 1 spot it will get stuck over and over again but move it on and it's fine again.

The default for max unsticks is 10 but you can change that in your profile options. The latest revision 572 has a new profile event. It is all about what happens when you reach the max unsticks. you can find the changelog here http://www.solarstrike.net/phpBB3/viewt ... =21&t=2198

Hope this helps in some way.

[kx9488]

I'd have to double check but from memory the unstick counter resets if you travel more then 3 waypoints.

I have noticed this issue of getting stuck for no apparent reason, as yet I haven't been able to pinpoint the actual cause. I don't use quick turn as you mentioned in your other topic. I notice it seems to happen in certain places though, bot will run smooth and then this 1 spot it will get stuck over and over again but move it on and it's fine again.

The default for max unsticks is 10 but you can change that in your profile options. The latest revision 572 has a new profile event. It is all about what happens when you reach the max unsticks. you can find the changelog here http://www.solarstrike.net/phpBB3/viewt ... =21&t=2198

Hope this helps in some way.

ty,

so if i used the waypoint code you created in another topic could i insert it in there to help the unstick process like this?

Code: Select all

<onUnstickFailure><![CDATA[
          __WPL:setDirection(WPT_FORWARD)
   	 ]]></onUnstickFailure>

[rock5]

With my work and testing to improve RoMScript reliability, I found that it often just fails to press the macro hotkey. This has led me to think that maybe other keys sometimes fail to be pressed. Maybe in certain situations if the forward button doesn't get pressed it can get stuck adjusting the angle but not moving forward so it thinks it got stuck?

I've looked at the code but it's hard to follow. I think the section it might be getting stuck in is this.

Code: Select all

		if( angleDif > math.rad(15) ) then
			--keyboardRelease( settings.hotkeys.MOVE_FORWARD.key );
			--keyboardRelease( settings.hotkeys.MOVE_BACKWARD.key );

			if( angleDifference(angle, self.Direction + 0.01) < angleDif ) then
					keyboardRelease( settings.hotkeys.ROTATE_RIGHT.key );
					keyboardHold( settings.hotkeys.ROTATE_LEFT.key );
					yrest(100);
			else
					keyboardRelease( settings.hotkeys.ROTATE_LEFT.key );
					keyboardHold( settings.hotkeys.ROTATE_RIGHT.key );
					yrest(100);
			end

This is the only section that doesn't try to press the forward key again. Maybe this section only works if you are moving forward. Maybe you could try adding

Code: Select all

			keyboardHold( settings.hotkeys.MOVE_FORWARD.key );

just after it so it starts moving forward again.

[kx9488]

With my work and testing to improve RoMScript reliability, I found that it often just fails to press the macro hotkey. This has led me to think that maybe other keys sometimes fail to be pressed. Maybe in certain situations if the forward button doesn't get pressed it can get stuck adjusting the angle but not moving forward so it thinks it got stuck?

I've looked at the code but it's hard to follow. I think the section it might be getting stuck in is this.

Code: Select all

		if( angleDif > math.rad(15) ) then
			--keyboardRelease( settings.hotkeys.MOVE_FORWARD.key );
			--keyboardRelease( settings.hotkeys.MOVE_BACKWARD.key );

			if( angleDifference(angle, self.Direction + 0.01) < angleDif ) then
					keyboardRelease( settings.hotkeys.ROTATE_RIGHT.key );
					keyboardHold( settings.hotkeys.ROTATE_LEFT.key );
					yrest(100);
			else
					keyboardRelease( settings.hotkeys.ROTATE_LEFT.key );
					keyboardHold( settings.hotkeys.ROTATE_RIGHT.key );
					yrest(100);
			end

This is the only section that doesn't try to press the forward key again. Maybe this section only works if you are moving forward. Maybe you could try adding

Code: Select all

			keyboardHold( settings.hotkeys.MOVE_FORWARD.key );

just after it so it starts moving forward again.

ty,

but im kind of dumb, where would i stick that code just paste it after all of those if/then lines like this so after it executes as it is sappose to it will move forward?

Code: Select all

		if( angleDif > math.rad(15) ) then
			--keyboardRelease( settings.hotkeys.MOVE_FORWARD.key );
			--keyboardRelease( settings.hotkeys.MOVE_BACKWARD.key );

			if( angleDifference(angle, self.Direction + 0.01) < angleDif ) then
					keyboardRelease( settings.hotkeys.ROTATE_RIGHT.key );
					keyboardHold( settings.hotkeys.ROTATE_LEFT.key );
					yrest(100);
			else
					keyboardRelease( settings.hotkeys.ROTATE_LEFT.key );
					keyboardHold( settings.hotkeys.ROTATE_RIGHT.key );
					yrest(100);
			end
keyboardHold( settings.hotkeys.MOVE_FORWARD.key );

and would this work in a profile the same as it would in a waypoint file?

Code: Select all

<onUnstickFailure><![CDATA[
          __WPL:setDirection(WPT_FORWARD) --to force move to next waypoint in numerical order after unstick fail
       ]]></onUnstickFailure>

thanks again

[lisa]

Code: Select all

<onUnstickFailure><![CDATA[
          __WPL:setDirection(WPT_FORWARD)
   	 ]]></onUnstickFailure>

This only makes the direction of the order of waypoints used forward, ie 1 2 3 so it wouldn't actually tell the bot to move in any way.
I think rock5 was saying to try something like this

Code: Select all

      if( angleDif > math.rad(15) ) then
         --keyboardRelease( settings.hotkeys.MOVE_FORWARD.key );
         --keyboardRelease( settings.hotkeys.MOVE_BACKWARD.key );

         if( angleDifference(angle, self.Direction + 0.01) < angleDif ) then
keyboardHold( settings.hotkeys.MOVE_FORWARD.key );
               keyboardRelease( settings.hotkeys.ROTATE_RIGHT.key );
               keyboardHold( settings.hotkeys.ROTATE_LEFT.key );
               yrest(100);
         else
keyboardHold( settings.hotkeys.MOVE_FORWARD.key );
               keyboardRelease( settings.hotkeys.ROTATE_LEFT.key );
               keyboardHold( settings.hotkeys.ROTATE_RIGHT.key );
               yrest(100);
         end

Something along those lines anywa, I think I'll do some testing.

[rock5]

Code: Select all

		if( angleDif > math.rad(15) ) then
			--keyboardRelease( settings.hotkeys.MOVE_FORWARD.key );
			--keyboardRelease( settings.hotkeys.MOVE_BACKWARD.key );

			if( angleDifference(angle, self.Direction + 0.01) < angleDif ) then
					keyboardRelease( settings.hotkeys.ROTATE_RIGHT.key );
					keyboardHold( settings.hotkeys.ROTATE_LEFT.key );
					yrest(100);
			else
					keyboardRelease( settings.hotkeys.ROTATE_LEFT.key );
					keyboardHold( settings.hotkeys.ROTATE_RIGHT.key );
					yrest(100);
			end
keyboardHold( settings.hotkeys.MOVE_FORWARD.key );

Yes exactly there.

and would this work in a profile the same as it would in a waypoint file?

Code: Select all

<onUnstickFailure><![CDATA[
          __WPL:setDirection(WPT_FORWARD) --to force move to next waypoint in numerical order after unstick fail
       ]]></onUnstickFailure>

The command should work but I don't know if it would fix your problem.

[kx9488]

ty,

im using both codes. i think the situation might have improved, but i'm not sure by how much... or if it's just me wanting it to be better.

it seems like i don't enter the unstick trials as much, but i have caught it doing it a few times still...

is there anyway to edit out code where it tells it to run around in circles durring the unstick trial. i'd like to remove that and see if i can stay botting all night.

if someone wants to record me running in loop-d-loops though, its going to look pretty bad.

im not worried about it getting stuck and having to run around things anyway... it would either fall off a ledge or bot normally

***edit***

i also notice that i am moving to waypoints in squiggle S-like patterns sometimes now where i might have been stuck running in circles from unstick trials.

[rock5]

I'm not aware of unstick code that makes you run in circles. That's another problem. My fix was only for the swaying left and right problem. Did you see anymore the swaying left and right problem?

[kx9488]

I'm not aware of unstick code that makes you run in circles. That's another problem. My fix was only for the swaying left and right problem. Did you see anymore the swaying left and right problem?

i think what i'm noticing now, the S-shaped paths here and there are where it would be running in circles. i think normally it would stop and start panning left and right then entering unstick (running in circles). now it stops and moves left to right as it runs.... not every time, but about as much as it was doing the left to right thing.

everytime it runs unstick it does circles but it seems like it isn't doing it nearly as much as it was.

something improved... but the squiggly paths are only a little better then circles.

it definatly does unstick A LOT less... but something is funky, making it turn a lot when it moves

***edit***

i think it mostly happens when there is nothing to attack.

it moves to the waypoint. and if it kills everything before it touches the waypoint it then tries to find the waypoint it was sapposed to move to. it could have already passed over the waypoint it is trying to find, so it searches for it. when it runs in circles i think it is trying to find it... and with the new code telling it to move on, i think it skips it instead of getting stuck in cirlces and does a wavy S-pattern instead of staying in 1 spot and unstick trialing out.

[rock5]

Like I've said previously, I don't 100 % follow the move function. Maybe putting that 'move forward' keypress there was a mistake.

I think it turns in circles because it can't turn fast enough to face the waypoint while it's moving.

Maybe the problem is just that it's turning too much? I think what determines the amount it turns is those 'yrest(100)'s. You could try reducing those. I don't know why they would be wrong though. I'm sure they would have been tested to see if they turned the right amount.

[kx9488]

I think I'll do some testing.

how did this work out?

.
Maybe the problem is just that it's turning too much? I think what determines the amount it turns is those 'yrest(100)'s. You could try reducing those.

i suppose i could, there are like 25 of those phrases in the player.lua though , change all of them at the same time or just phrases close to lines like these

( angleDif > math.rad(15) )
&
"-- to prevent unsticking"
&
( angleDifference(angle, self.Direction + 0.01) < angleDif )

stuff like that?

P.S.
am i the only one experiencing stuff like this? lisa mentioned she did. is this mostly a problem for melee chars? any other members with ranged skills have trouble like this?

and what versions are people running? i am now using only these files:
https://micromacro.googlecode.com/svn/trunk
that have been updated and reverted with the newset version of TortoiseSVN downloaded from here:
http://tortoisesvn.net/downloads.html

the only modifications i have are short bits of code in the player.lua at line #2003 after this line:

-- Attempt to unstick the player
function CPlayer:unstick()

line #2003:

Code: Select all

-- after 1x unsuccesfull unsticks try to reach nearest waypoint
   if( self.Unstick_counter == 2 ) then
   __WPL:setWaypointIndex(__WPL:getNearestWaypoint(player.X, player.Z, player.Y));
   end;

and this at line #1897 in the player.lua after these lines of code

if( angleDif > math.rad(15) ) then
--keyboardRelease( settings.hotkeys.MOVE_FORWARD.key );
--keyboardRelease( settings.hotkeys.MOVE_BACKWARD.key );

if( angleDifference(angle, self.Direction + 0.01) < angleDif ) then
keyboardRelease( settings.hotkeys.ROTATE_RIGHT.key );
keyboardHold( settings.hotkeys.ROTATE_LEFT.key );
yrest(100);
else
keyboardRelease( settings.hotkeys.ROTATE_LEFT.key );
keyboardHold( settings.hotkeys.ROTATE_RIGHT.key );
yrest(100);
end:

line #1897:

Code: Select all

keyboardHold( settings.hotkeys.MOVE_FORWARD.key );

This bit of code in my profile:

Code: Select all

<onUnstickFailure><![CDATA[
          __WPL:setDirection(WPT_FORWARD)
       ]]></onUnstickFailure>

my waypoint files are pretty much 100% generated from the function in micromacro:
"rom\createpath.lua"

and my profile settings:

<!-- Combat options -->
<option name="COMBAT_TYPE" value="melee" />
<option name="COMBAT_RANGED_PULL" value="false" />
<option name="COMBAT_DISTANCE" value="50" />
<option name="MAX_FIGHT_TIME" value="1" />
<option name="DOT_PERCENT" value="90" />
<option name="ANTI_KS" value="true" />
<option name="MAX_TARGET_DIST" value="100" />
<!-- Waypoint and movement settings -->
<option name="PATH_TYPE" value="waypoints" />
<option name="WANDER_RADIUS" value="500" />
<option name="WAYPOINT_DEVIATION" value="0" />
<option name="QUICK_TURN" value="false" />
<!-- Loot settings -->
<option name="LOOT" value="true" />
<option name="LOOT_IN_COMBAT" value="true" />
<option name="LOOT_DISTANCE" value="100" />
<option name="LOOT_PAUSE_AFTER" value="0" />
<skills_rogue>
<skill name="SCOUT_VAMPIRE_ARROWS" modifier="" hotkey="VK_2" priority="70" />
<skill name="ROGUE_SHADOWSTAB" modifier="" hotkey="VK_4" priority="90" />
<skill name="ROGUE_LOW_BLOW" modifier="" hotkey="VK_5" priority="80" />
<skill name="ROGUE_WOUND_ATTACK" modifier="" hotkey="VK_6" priority="60" />
<skill name="ROGUE_COMBAT_MASTER" modifier="" hotkey="VK_8" Priority="100" />
</skills_rogue>

those are all of the deviations from the stock files i can think of. my waypoints could be placed in a conflicting manner is the only thing i could think of.
anyone want to discus ideal waypoint placement scenarios? i try to put them in straight lines, and in the middle of mob clusters... and if they double back close to eachother i try to make the entry side go up one side and the exit side go to the other side as to seperate them.

or do all of you have vastly customized player.lua's among other files?

-ty

[rock5]

I just noticed that if you have quickturn enabled and it quickturns, it doesn't do a player:update() to update it's direction.

Try adding a "self:update()" after this around line 1852;

Code: Select all

		-- Continue to make sure we're facing the right direction
		if( settings.profile.options.QUICK_TURN and angleDif > math.rad(1) ) then
			self:faceDirection(angle);
			camera:setRotation(angle);
		end

[kx9488]

I just noticed that if you have quickturn enabled and it quickturns, it doesn't do a player:update() to update it's direction.

Try adding a "self:update()" after this around line 1852;

Code: Select all

		-- Continue to make sure we're facing the right direction
		if( settings.profile.options.QUICK_TURN and angleDif > math.rad(1) ) then
			self:faceDirection(angle);
			camera:setRotation(angle);
		end

thank you rock,

line #'s 1848-1859 look like this:

#1848: -- We're still making progress
#1849: if( dist < lastDist ) then
#1850: self.LastDistImprove = os.time();
#1851: lastDist = dist;
#1852: elseif( dist > lastDist + 40 ) then
#1853:
#1854: -- Make sure we didn't pass it up
#1855: printf(language[29]);
#1856: success = false;
#1857: failreason = WF_DIST;
#1858: break;
#1859: end;

so should i make it after line 1859? or slap it under 1852? i know there is a nice convenient place right were you said but i want to make sure because the "end's" and the notations it would be around were kind of ambiguous. the chronological order of: "-- Continue to make sure we're facing the right direction" should it go after,

"-- We're still making progress"
or
"-- Make sure we didn't pass it up" ?

-ty

[SirPacco]

I've exerienced the same issues, but only in Sascilia Steps, in other places the bot run smoothly.

[rock5]

so should i make it after line 1859? or slap it under 1852? i know there is a nice convenient place right were you said but i want to make sure because the "end's" and the notations it would be around were kind of ambiguous. the chronological order of: "-- Continue to make sure we're facing the right direction" should it go after,

"-- We're still making progress"
or
"-- Make sure we didn't pass it up" ?

-ty

No, after the code I said above. Obviously I gave you the wrong line number. You should insert it around line 1883.

[kx9488]

rofl... i wasn't trying to be a wise guy i was going to actually paste this code in:

Code: Select all

-- Continue to make sure we're facing the right direction
      if( settings.profile.options.QUICK_TURN and angleDif > math.rad(1) ) then
         self:faceDirection(angle);
         camera:setRotation(angle);
      end

i just realised this is what i was adding:

Code: Select all

self:update()

DOH!

thank you,

ill try it out later, i needed to take a break on my main, i think people were reporting me so i stopped on him and started doing miller's ranch.

they seem to unstick timer out too... ill try pasting it in for them

possibly from lagg though im not positive (running 3 at a time)

[Trivia69]

I too am experiencing the same issue. As I searched several forums before finding this post, there are others that are experiencing the same problems.

After initial installation of the bot it was working fine and I had some issues with a particular waypoint file and then another. I then deleted all the files in the rom directory and did an update via SVN and I started to have this issue. What I don't understand is there wasn't an update to the server nor was there an update via SVN that I know of in the time frame I did the deletion, it was just a clean installation of the files.

I just did an SVN update and noticed there was a file that had a date change to 2/7 from a previous date in the ingamefunctions folder. I also just copied those two files into my addons folder again. I will run the bot today with a few newly created waypoint files and test to see if different amounts of waypoints have an effect on it.

Also I did have a waypoint variance adjusted to 0 and I'm going to adjust that to a given number and see if that adjusts anything if it continues to try to unstick the non-stuck

[SirPacco]

I really want to help in solving this question but I can't know how.
Can someone of the Solar strike programmer tell me what they need me to test or feedback on this problem?
The problem can be easly reproduced trying to bot in sascilla steps.
First I thought that it was a problem on the y axis, but also in the plain part the problem is there
After i thought it was a problem because the high respawn rate of the mobs, but even in low respawn zone the problem is still there, the pg stuck and start to face right/left for a couple of tries and then do the unstuck procedure.
The strange thing is that the pg start a movement through the waypoint, like
2
3
4
unstuck
go to 2
unstuck
etc etc

[rock5]

I really want to help in solving this question but I can't know how.
Can someone of the Solar strike programmer tell me what they need me to test or feedback on this problem?
The problem can be easly reproduced trying to bot in sascilla steps.
First I thought that it was a problem on the y axis, but also in the plain part the problem is there
After i thought it was a problem because the high respawn rate of the mobs, but even in low respawn zone the problem is still there, the pg stuck and start to face right/left for a couple of tries and then do the unstuck procedure.
The strange thing is that the pg start a movement through the waypoint, like
2
3
4
unstuck
go to 2
unstuck
etc etc

Can you send me a copy of your waypoint file and profile so i can test it? If you don't want to post them on the forum, send them to me via PM. If I can reproduce the error, maybe I can fix it. Does it get stuck in a particular place or places or does it get stuck anywhere?

By the way, what version of the bot are you running? I made an attempt to fix this problem in revision 577.