Robo Home | RoboRumble | Changes | Preferences | AllPages

Difference (from prior major revision) (no other diffs)

Added: 6a7,9
It's happened to me before, not sure why. I made the comment so that we all double-check the list -- the rumble client crashes if there's a blank line (lots of text formatting glitches crash this poor program) so no battles go through until it's fixed. --Darkcanuck

Skilgannon what browser are you using? After Darkcanuck's last edit to the participants list, where he said someone's browser likes to add blank lines, I thought it was me because I'm using Chrome, and since it is relatively new it probably has some bugs. So I started comparing all revisions and saw that after every update of DrussGT the is a remove empty line edit, and the weirdest thing is that you always modify line 306/307 and an empty line is always added to line 202. --Zyx

Yeah, very strange. I'm using Firefox, running on Ubuntu 8.10. This time, when adding the new DrussGT, I checked, and when I loaded the edit page it had the empty line already in it. I'm not sure what's going on... it's the same browser that every other user of the latest Ubuntu is using... maybe my router can't handle certain character sequences or something? I'm stumped. -- Skilgannon

I've check all special charactors but there are no difference from other lines! All lines end with \r\n (as from my PHP code that perform a HTTP request to server and print hex code out) --Nat

It's happened to me before, not sure why. I made the comment so that we all double-check the list -- the rumble client crashes if there's a blank line (lots of text formatting glitches crash this poor program) so no battles go through until it's fixed. --Darkcanuck

I added my bot to the list yesterday, but it's not on the ranking list yet. How long it used to take to add a new bot? (Robot name:TieFighter) -- HVilela

OK, I saw that when RoboRumble downloads my Robot it says: "Downloaded file is wrong or corrupted". I can manually download and run it, so what it means? -- HVilela

Well the properties file in you jar says "robot.classname=henriquevilela.HenriqueMasterRobot?" and your class is named is henriquevilela.TieFighter, that is probably the problem. You should package with robocode, open your robocode client, go to Robots -> Package your robot for upload, select your bot and fill in the info. - Zyx

Sorry, I'm literally on my way to bed, but wanted to post that mbh.Mbh is crashing every time, getting a 0 score. My bet is on it being compiled with Java 6. I'll look tomorrow if nobody has confirmed / denied this... -- Voidious

I just updated to Java 6, and it's running fine on my machine...probably what the problem is. -- Skilgannon

To the author of mbh.Mbh: could you please compile your tanks with Java 5 compatibility? Not all of us (especially us Mac users) are running Java 6, so a Java 6 bot crashes every time, gets a 0 score, which is discarded by the RoboRumble server, so running my client would just be wasting cycles endlessly. Also: welcome to the wiki, you should make yourself a page and tell us about yourself and your bots. =) -- Voidious

Nice to see people are still playing this game. I'll go take a look at my nanos again to see if they can be competative again :) --Miked0801

Also for ola.Puffin - could you compile for 1.5 compatibility, please? Definitely been wasting some cycles on my client, but I didn't realize because several bots are in need of battles. Modifying my client to skip certain bots is kind of a pain... -- Voidious

DrussGT is crashing on somebody's system...it's compiled with jikes ( therefore 1.5 compatible) and doesn't use the Rules class, so I'm stumped. Does anybody have any ideas? If possible, could you give me the error's stack trace? Thanks. -- Skilgannon

Hey, it'd be great if Bemo.* were compiled with Java 5 compatibility. I hate seeing my clients have been churning away on a bot without anything to show for it. Man, I hope Leopard has Java 6 good to go, as I'm really tired of making this request! -- Voidious

I have never seen this much RoboRumble computing power! Komarious is on pace to hit 2,000 battles in well under 2 hours here, and Bemo.Sweet30 is getting some battles in there, too. Too cool! -- Voidious

Komarious ... uh oh. Maybe I should take advantage of all the power right now, too ... -- Simonton

 *Flex* --David Alves

I released 4 different versions of DrussGT, and they all got to 2000 battles in about an hour. What system do you have running there? Some serious grunt there, man. -- Skilgannon

Ok, you guys all just seriously need to STOP having so much fun until I'm ready to join in! I have a decent gun now, but it's skipping turns, and I haven't even started on the surfing. Sheesh. On a different note, what's supposed to be the difference between /Chat and /ParticipantsChat? -- Simonton

This one is for participants. :-D -- Voidious

A local problem running the rumble. I change over to Moebius 2.2 and now whenever it attemps to report results, it shows DustBunny 3.2 fought instead then crashes with an array index out of bounds error. Any ideas? --Miked0801

With CC 2rho.01b, I just wanted another test of how much the new RoboRumble versions might drift the ratings upwards for new bots, because I'm just curious like that. =) [Dookious 1.554b vs 1.554] showed it to be about 5 points, but two tests is a lot more reliable than one. I don't think PEZ will mind me re-releasing a version of CassiusClay. -- Voidious

It looks like the modern RoboRumble versions give a 5-7 point rating advantage over the old ones, at least for bots in the 2060 - 2120 range. I wonder which bug fixes have made the most impact? -- Voidious

It might be the accurate locations of hit-bullets, making the surfing more accurate (5 - 7 points isn't much.. especially against all bots). That one looks most likely to me.. -- Skilgannon

I would also consider the "stopping robot, no score will be generated" thing, since that would only impact the last bot surviving. And this one might have been fixed too long ago to contribute to what you're seeing now, but the new algorithm for bullet-hit-bullet detection creates a lot more of those events, which tor bots use to learn GF guns without getting damaged in the process. -- Simonton

Until we get the repository replacement going: if you need web space to host your bots, feel free to e-mail them to me (see ContactInfo) and I'll upload them to my web space and send you back the URL. -- Voidious

Any one know where to find the id number that should be specified in the Participants list??? I have uploaded my bot on Robocode Repository but i have not found any id....

Thanks Voidious! i have found the id and uploaded my bot :)

Hey, are you other guys still getting errors when trying to upload battles? Last I checked, my RR client still got a "500 error" when trying to upload results. Those results are old, still saved / not uploaded from when this problem started, but I'm not sure if that matters. -- Voidious

I haven't been running a client....so no idea. -- Skilgannon

Yep, still having errors, indeed only old results. Only with some specific bots though, like Ugluk and/or BlackDestroyer. Maybe the results file of that bot is damaged in such a way that no updates for it are possible. Mind you that there are still a lot of resultfiles damaged (like Dookious, see details) and it seems that it is not automatically repaired. -- GrubbmGait

So I'm wondering....with these problems...should I bother releasing the new DrussGT? -- Skilgannon

I don't see how it will hurt anything if you do. As far as I know, the disk space problems on the server were resolved, but some bots have corrupted data files. I haven't had much chance to look into it with the holidays and vacation that just went by, but I'll try to look at the RR server code that Pulsar sent me soon and see if there's some way we can hack it to deal with the corruption... -- Voidious

New bots released will have no corruption. As for older bots, maybe it is enough to edit their datafiles and just delete the corrupted lines as they are plain text-files. -- GrubbmGait

What about removing ru.master.Hydra, its getting alot of zero results. ;( --Krabb

It also started getting a lot of zero results here too. =\ -- Rednaxela

I have bots that fail to download regularly, such as agd.Mooserwirt2, bayen.nut.Squirrel, winamp32.micro.MicroMacro?, etc. Frankly these generally just hang up my rumble client, forcing me to constantly watch it and restart it until it fails quickly enough to not just hang the downloader. This happens in both windows and linux with 1.5.4... --Chase-san

I never bothered to get those, I just run and get them from the net, which gets the very large majority, save maybe 10 or so that are just not there anymore, also the lag into the downloading is gone, ergo they don't hang anymore, donno why. Really they should all be kept up to date with somewhere thier jar can be gotten, and some I don't think are just temporarily down, such as MicroMacro?'s, I have seen that one every time I have ever started my client. --Chase-san

Anyone have the .jar for deewiant.Anomaly? its the only one I am missing. I'd like to make a new .zip for all the bots. --Baal

http://designnj.de/roboking/deewiant.Anomaly_0.1.jar there it is --Krabb

http://home.versatel.nl/gheijenk/robocode/things/deewiant.Anomaly_0.2.jar and here is version 0.2 -- GrubbmGait

Thanks, Guys! --Baal

One quick question that I'm wondering about RoboRumble. What happens if my bot uses onStatus() which to my understanding was introduced in 1.5.x, but people are still using 1.4.x on RoboRumble (cpu constant and such perhaps)? Would those clients ignore my bot which only works for RoboCode 1.5.x or would they give my bot bad scores? -- Rednaxela


Woah there! The latest DrussGT is outranking Dookious! I wonder if this will last. -- Rednaxela

So far it's slightly below Dookious in the comparison, but still many pairings to go. -- ABC

Quite a high difference, but true, still many matches to go yet. --Chase-san

The rating always seems to drop when fighting the last third of the opponents. It will be a close call though. I bet Voidious is already busy with the next version, expecting it before they reach 1500 battles. -- GrubbmGait

I should have released DrussiousGT using 1.0.5's movement instead of 1.0.7's, I think I messed something up in there. I think DrussGT will come in just below Dookious, but DrussiousGT a bit above. -- Skilgannon

Hmm, doesn't seem those predictions are holding too true. DrussGT just passed 1000 battles, and it's still just a little (1.5 points) above Dookious still, and DrussiousGT is trailing about 10 points behind either. EDIT: And it didn't quite last, as of 1232 battles DrussGT is about 1 point lower than Dookious. -- Rednaxela

Just a note about BlackSwans - I think we should rerelease it as 0.60b once we've gotten through the backlog of current bots that need 2,000 battles. I'd already deleted it from my rumble dir and I have blocked RobocodeRepository in my local "hosts" file, since I was having the same problem... -- Voidious

Huh... I'm getting this error when trying to run the RR@H client now... Is there some kind of corruption of the results file? I'm suddenly getting this on two separate computers when trying to run the rumble after clearing out the temporary files (roborumble/files/* roborumble/temp/* and robots/.robotcache/*)

Executing battles ...
Fighting battle 0 ... jk.Waylander 0.3.7,pez.rumble.CassiusClay 2rho.01b
Exception in thread "Application Thread" java.lang.ArrayIndexOutOfBoundsException: 0
	at roborumble.battlesengine.BattlesRunner.runBattles(Unknown Source)
	at roborumble.RoboRumbleAtHome.main(Unknown Source)
-- Rednaxela

You also have to remove robots/robot.database so that it knows to rebuild robots/.robotcache. -- Skilgannon

Ah, thanks. Got two computers back on running the rumble now (while a third is on RoboResearch like often lately ;)). Thanks -- Rednaxela

What the heck? Some results got uploaded showing RougeDC losing to bots it gets 98+% against... Bad client? Or perhaps I'm the CPU constant is set low there? Gah... -- Rednaxela

Yikes! I think that was me... I was running a torrent download on the same machine as my one client, and the torrent client suddenly decided it would be a good idea to allocate itself about a gig of ram. Needless to say the system ground to a halt, probably causing RougeDC to skip a whole bunch of turns. I lost all the ssh output (it required a hard reboot), but that was about half an hour ago. Sorry bout that. It seems about 3 bots, huh? FloodHT, Abot, Hawkwing. -- Skilgannon

Yeah, those appear to be the ones (and I see they were indeed from your client, looking in the results file). I'll do a Gamma4 release in a moment to replace it which will be the same except with a few tweaks that seem to have proven themselves in testing. -- Rednaxela

I get this error: Iteration number 0 Downloading participants list ... Exception in thread "main" java.lang.StringIndexOutOfBoundsException?: String index out of range: -1

	at java.lang.String.substring(String.java:1949)
	at roborumble.netengine.BotsDownload?.downloadParticipantsList?(Unknown Source)
	at roborumble.RoboRumbleAtHome?.main(Unknown Source)

That's because of some blank lines Skinn3r added in RoboRumble/Participants by mistake. I just fixed that now and it should work. -- Rednaxela

ok, now my client work. But still have one problem: i can't dowload any bot from robocoderepository, i get the error: Could not connect to http://robocoderepository.com/Controller.jsp?submitAction=downloadClass&id=XXXX and for my bot: Could not find asd.Cthulhu 1.0 from http://robocoderepository.com/Controller.jsp?submitAction=downloadClass&id=3416 also the link dosn't works... --[Asdasd]

This this is very strange. The same thing happens here, and the URL is exactly the same as one that does work on robocoderepository.com. From what I can tell, robocoderepository.com has for some strange reason started blocking downloads when the download URL is visited from anywhere but itself... -- Rednaxela

I've tried to import manually my bot after adding it in the participants list but roborumble try always to download it... Not a real problem, I will do my bot better^^

I've noticed in the PL DrussGT and Shadow are missing some battles, despite having filled their 2000. It probably has something to do with bots that were lost in the repository crash. Any ideas? I'm too tired to think right now... -- Skilgannon

Sorry 'bout the blank lines, donno how they got there. Since I only edited the line for my bot. :( Skinn3r

I just noticed that the roborumble rankings seem to have lost a lot of bots. There are only 431 right now in the list, and the total battles is 3.7 million. There should be about 640 bots, and I'm sure there used to be more than 5 million battles. The participants list still has 640-ish entries. Any one understand what's happened here? -- Tim Foden

In the past, this happens if someone puts the wrong participants URL in their RoboRumble config; if the participants page itself had some deleted, that would do it too. The good news is that once the problem is fixed, when a bot gets one more battle, it will reappear with all of its previous battles - so it won't take weeks of RR processing to get the rankings right again. I'm at work and don't have time to look into it right this second, but could everyone check their RoboRumble config, especially anyone who just set one up? -- Voidious

This time I'm OK^^, I'm running the client and rankings are right --Asdasd

I just noticed that someone removed Moron due to it winning battles it shouldn't. I think that isn't a good fix and that Moron should be put back, because it sounds like a bad client to me. Thoughts? -- Rednaxela

Moron's author probably doesn't know about the bad clients problem. If you're reading this check this server out: http://abchome.aclsi.pt:8080/rumble/RatingDetails?game=roborumble&name=pedersen.Moron%202.0, Moron has no wins and 0.06 average percent score. -- ABC

Finally finished running approx 280 'repair' battles for ad.last.Bottom. (running 6 battles against each opponent to reduce the influence of bad results with a factor 8) Not something I would recommend to anyone though, as it took me over 3 weeks, about 25 to 30 hours real processing. Point is that it will take loooong to repair all bad results for all bots. Would it be worthwile to process the top-25 and bottom-20 or so to have a reasonable reliable (for those bots) ranking/rating, or would it be just a waste of my cpu-cycles. I'll see if I can automate it, so I can run it unattended, but it would still take at least one day per bot. -- GrubbmGait

I believe Pulsar's server is beyond repair, that's why I started my own from scratch. In 3 weeks we can probably reconstruct a full pairings ranking, and it's been longer than that we have a messed up ranking. -- ABC

If the author of jsal.Jsalbot 1.0 is reading this. I'm not 100% sure, but I think you have your jar and package as "js", then you saw (my guess) someone who already used js.* so you changed it to jsal.* on the participants list, you should rename your package to jsal, and then repackage the bot so the jar name changes too. Right now my client says your jar is invalid, but I thinks that's all. -- Zyx

I am the author of jsal.Jsalbot 1.0, I did see someone already used js, so I changed my package to jsal. However, I changed the package of my robot and re-exported it. I think I might have not reuploaded it though. I will fix that and see if that helps. Thanks -- jeremybub?

No problem, I saw your bot already fought battles :). Hopefully you already now, but if not there is a new wiki (http://testwiki.roborumble.org), you should check it out. Good luck with your robocoding -- Zyx

Robo Home | RoboRumble | Changes | Preferences | AllPages
Edit text of this page | View other revisions
Last edited March 25, 2009 7:42 EST by Darkcanuck (diff)