#Eulora log for Tuesday, 2017-02-21

Latest Log | Search | Help | Eulora Wiki (eulorum.org)

« Monday, 2017-02-20
Or choose a date:
Wednesday, 2017-02-22 »

Birdmanim also making bundles to have tinkered by someone better00:10
Birdmanjust got done rather00:10
lobbesbotAUCTION # 25: 200 inconsequential sacrifice tokens q100 Opening: 90mn Highest Bid: No bids Ending: 2017-02-21 02:46:38 UTC (2 hours 28 mins)00:18
lobbesbotAUCTION # 25: 200 inconsequential sacrifice tokens q100 Opening: 90mn Highest Bid: No bids Ending: 2017-02-21 02:46:38 UTC (1 hours 28 mins)01:18
lobbesbotAUCTION # 25: 200 inconsequential sacrifice tokens q100 Opening: 90mn Highest Bid: No bids Ending: 2017-02-21 02:46:38 UTC (0 hours 28 mins)02:18
lobbesbotAUCTION # 25 has ENDED: No Sale. Attn: hanbot02:47
mircea_popescusomething tells me i'ma regret this03:52
mircea_popescubut... no cash.03:52
diana_comanthis update totally ruined the pastures for my microgoats, grrr07:59
diana_comanI got a microgroat q6 out of a small, not even worth the cft that went in basically07:59
diana_comanand I can confirm that somehow claims themselves influence output q: I got different qualities from tiny claims too, same resource, same bundle q08:07
diana_comanben_vulpes, I've put together a short example of cal3d model + use with cs to have as starting point; the model there is provided by cal3d themselves so it's exactly the sort of output we want from a blender exporter: all the files in folder cally + test.cal3d (the description of the model in a format that cs can use); source code and brief description here: http://dianacoman.com/Eulora/sources/sources.html15:54
diana_comanand anyone else interested in models/graphics for eulora (hanbot ?) ^15:55
hanbotyeh i'll definitely have a look16:23
ben_vulpesdiana_coman: for your calibration, i'm testing mod6's v.pl atm. will look into the blender project once i deliver tests to him16:28
diana_comankk, no rush16:28
diana_comanhanbot, if you have time and want to give it a try, there is apparently some cal3d exporter for blender 2.62 - no idea if it works at all or is of any use whatsoever https://wiki.blender.org/index.php/Extensions:2.6/Py/Scripts/Import-Export/Cal3D_File_Format(.cfg)17:41
diana_coman(I would expect it to work with 2.66, if it works at all)17:42
thestringpulleri got this one to work: https://github.com/alexeyd/blender2cal3d last night; outputs (cmf/caf/csf/xrf). but it's simple and you had to export one armature at a time and merge the cfg files19:17
*** Quits: diana_coman (~diana_com@unaffiliated/diana-coman/x-8319396) (Quit: Leaving)19:20
Birdmanso what is to be done with the mining bot, maybe have it use multiple containers to not mix qualities? wouldn't work eventually with getting overladen19:21
danielpbarroni want mine to just mix on the fly19:25
danielpbarronand probably switch to a must-have-bundle method19:26
danielpbarronseems like the easiest way to control how many threads to use on each different thing19:27
danielpbarronnot really "how many threads" anymore19:27
Birdmanpre make bundles19:27
danielpbarronand let it build ords too, if the bundle is there19:28
Birdmanits also not giving the build any time for me, had to go manually build smalls that the bot made bundles in and left19:28
Birdmanstill, you'll eventually just lose tons of value with it mixing as it foes19:28
danielpbarronwhen i say mix on the fly i mean carefully19:29
danielpbarronlossless mix19:29
Birdmanmaybe have it mix qualities that are a certain amount close to each other19:29
Birdmanq0-50 mixes, 51-100 etc19:29
Birdmanfor now im just going to barehand a few thousand times and build smalls manually19:31
danielpbarronand the method is simply: put all different quality piles into container, pickup container, match piles into pairs where quantity is same and quality is different but same even/odd. mix paired piles, if results have duplicate quality: merge them, repeat until there are only 2 piles; put them back into container and continue exploring19:31
danielpbarronthe matching algorithm is also simple: separate piles into even side and odd side. on each side, take the 2nd largest pile and and equal amount from the largest pile as first pair, repeat until one pile left that cannot be paired. mix all the pairs. sort back into even odd, repeat19:33
danielpbarronone exception, when the 2nd largest pile is greater than 4999, have to only grab 499919:33
danielpbarronthe pairing step is a literal step. you cannot mix a pile with a picked-up part of another pile if the two add up to more than 999919:34
danielpbarroneven though the resulting mixed pile isn't greater than 999919:35
Birdmanthat'd still lose you quality over time19:58
*** Joins: diana_coman (~diana_com@unaffiliated/diana-coman/x-8319396)19:58
Birdmanehh, interestingly enough my mining tinies take 1-3 bits but the bot still gets the "you dont know what to do with this" when it tries to build20:03
Birdmanwhich is a good thing for now20:03
Birdmanmaybe because it is trying to use the bit, not combine it into a bunfler20:04
diana_comanhm, it shouldn't; Birdman, mind testing that manually and reporting it properly? meaning: 1 bundle+use for enum with 1-3; does it work? if not, what enum is that, etc20:33
Birdmansure one sec20:33
Birdmanpss tiny enum, instructions are 1-3 lbn combined to pss tiny exploration bundle to extract. put one lbn into claim click use get the message "you dont know what to do with this." Clicking combine actually gives me the same error message20:35
Birdmancan a bundle not be comprised of one thing?20:36
diana_comanit might be that particular bundle that is borked; will check; thanks20:36
diana_comanftr use/combine are interchangeable now; there shouldn't be any more silly errors that oh no, it was combine, not use or the other way around20:39
diana_comangot it, a bit of a corner case; will be fixed this Wednesday; thanks Birdman for catching that20:46
diana_comandanielpbarron, that sounds reasonable as an algo - you implementing it?21:07
danielpbarroni will eventually if nobody else has by the time i'm painted into a corner21:23
danielpbarronand no Birdman it doesn't lose quality21:24
danielpbarronwhen you mix equal amounts of even quality or odd quality you get an even division to an integer quality and no loss21:24
danielpbarronif you add two odd numbers you get even, add two even you get even. even number is divisible by 2 (and so is your final quantity because it is the result of combining two equal sized piles)21:25
danielpbarronyou can do a lossless mix for as long as you have 3 piles of different quality21:25
danielpbarronwhen you finally have 1 odd pile and 1 even, you can't mix further unless they just so happen to combine in such a way that it similarly does an even rounding, but that's a rare edge case probably not worth complicating the code over21:27
danielpbarronmost the things in my storage are kept this way, and odd pile and an even pile21:27
diana_comanwhy wait to be painted into a corner though danielpbarron ?21:30
danielpbarronthat's usually what gets me coding something21:31
DicePowerIs installing foxybot mandatory now-a-days?21:41
mircea_popescunot really.21:42
mircea_popescusignificasntly less useful since changes less wed, until someone patches it.21:42
mircea_popescuwould you like to finally do something useful and patch it ?21:43
DicePowerI'm getting fatal linking errors trying to compile Eulora's source code related to foxybot, even though I didn't intentionally install it.21:44
mircea_popescu9>c:\eulora-v0.1.2b\eulorav0.1.2\src\client\foxybot\botactivity.h(91): warning C4305: 'argument' : truncation from 'double' to 'float' << you did include it.21:45
mircea_popescubut it's entirely unclear why it wouldn't link, tbh.21:46
DicePowerWell, I got my tar.gz from http://minigame.bz/eulora/source/ but didn't explicitly install foxybot separately.21:46
DicePowerThe only thing I've really gone outside the bounds of the normal compilation process with was the stdint.h/inttypes.h thing.  I had to play around with those files a bit to remove some new errors in 0.1.221:49
DicePowerBut these linker errors sound unrelated.21:49
mircea_popescuit was bundled with 1.2b yes21:51
DicePowerAhhh, okay, that explains it.21:51
mircea_popescuyou can in principle just delete it (the folder i mean) and still work ; but there's at least one unrelated line in there (the topspops thing).21:51
mircea_popescuincidentally this sort of issue (sudden linker errors) sounds VERY much like a type mismatch.21:52
mircea_popescuso it probably is related.21:52
mircea_popescunow, eulora DID move away from broken ints to sane ints during migration to 0.1.2, and i don't know anyone cared enohgh to extensively debug the insanity that is microsoft. so you might have found something. what exactly were those new errors ?21:53
DicePowerI see21:53
DicePowerThey were related to recognizing int8_t and SCNd64 as keywords.  I tried many things so I can't recall the precise series of steps that worked to fix it, but I downloaded stdint.h and inttypes.h first and moved them around/edited them a bit (i.e., it appeared that the #if command in which SCNd64 was defined wasn't seen as true, so I added a SCNd64 definition outside of the #if.22:00
DicePowerI'm not sure that I actually ended up using stdint.h.  I think the real problem there was that the files in apppsclient are supposed to include psstdint.h, but there are two psstdint.h's and they were including they wrong one, or at least I thought they were.22:01
DicePowerOne psstdint.h has actual code, while the other is a one-liner that includes stdint.h.22:02
mircea_popescuwell, you gotta keep a good record of steps you took else this happens and is undebuggable.22:03
DicePowerWhich is why I downloaded stdint.h, but that ended up yielding redefinition errors, and I'm pretty sure I just changed everything to use psstdint.h and deleted the psstdint.h one-liner file so that the compiler could only find the one I wanted it to.22:04
mircea_popescumy advice is, drop everything you did so far ; mark your efforts as UTTER WASTE both by hand on paper and in your own head so you never engage in the sort ofg insanity again as long as you live22:04
mircea_popescuthen download the proper version, compile it and report what problem you run into.22:04
DicePowerWhat do you mean by "proper version"?22:05
mircea_popescuwhat minigame publishes as the code for eulora.22:05
DicePowerThe 0.1.2b tar.gz that I used, so just start over with the same file and keep better track of the steps I take?22:06
mircea_popescunot better.22:06
mircea_popescukeep perfect track of steps ; and don't even take steps you don't understand in the first place.22:06
mircea_popescuthis isn't a spicy jungle adventure, it's engineering. the pinnacle of reason-based human activity.22:06
DicePowerHmm, interesting22:07
mircea_popescuat all junctures gotta be able to say what you did and why you did it or else don't do it.22:07
DicePowerYa know, this nook of the programming community is very different than the rest of it, in a good way I think.22:09
mircea_popescui suspect so.22:09
DicePowerIn many other chat rooms, one gets labeled as lazy for asking a bunch of theoretical questions instead of jumping in head first and experimenting :P22:10
DicePowerI will do what you said and redownload the tar.gz and start again, keeping notes of every step.22:16
danielpbarronwhich one is DicePower ? the one who spent forever making new characters or the one who modified his client on windows? or was that the same guy?22:19
DicePowersame guy :D22:21
mircea_popescui think he might be the guy that kept making characters22:21
DicePowerI'm modifying the client to make characters on its own and record the stats of each character.22:21
danielpbarronDicePower, throw away all that, get the vanilla client running, and do some work in-game where it is needed and where you are best suited to show signs of utility. you can't possibly know how to modify the client or analyze the results of your science without having actually played22:42
danielpbarronyou already missed out on hundreds of thousands of coppers worth of skill items22:44
danielpbarronand you won't miss out on experience if that's your fear in committing to a character. the work we need you to do involves specifically having you ~not~ level up22:45

« Monday, 2017-02-20
Top o' the log
Wednesday, 2017-02-22 »

Generated using logs2html.py 2.14.0

#Eulora log-o-tron maintained by lobbes