00:59:29 Unstable branch on CRAWL.XTAHUA.COM updated to: 0.21-a0-405-gf2af9a1 (34) 01:05:47 Rast: "Improvements that are so obvious someone would already have thought of them if they made sense" 01:06:34 i feel bad for lasty that the thread has been sticking around for so long near the top of the forum 01:11:21 it's a good reminder for some useful minor fixes! 01:13:50 oh, is that the thread that made lasty show up and add the text to ?/gru about piety increasing the strength of all ru powers? i approve 01:15:18 I thought everyone already knew that about Ru 01:30:34 everyone who reads learndb knew, but it should have been in game 02:06:13 -!- amalloy is now known as amalloy_ 02:19:32 Unstable branch on crawl.develz.org updated to: 0.21-a0-405-gf2af9a1 (34) 02:29:08 -!- amalloy_ is now known as amalloy 02:58:12 Windows builds of master branch on crawl.develz.org updated to: 0.21-a0-405-gf2af9a1 03:11:34 Unstable branch on crawl.beRotato.org updated to: 0.21-a0-405-gf2af9a1 (34) 03:31:06 -!- amalloy is now known as amalloy_ 03:53:28 Monster database of master branch on crawl.develz.org updated to: 0.21-a0-405-gf2af9a1 06:47:09 New branch created: pull/644 (1 commit) 13https://github.com/crawl/crawl/pull/644 06:47:09 03Yermak02 07https://github.com/crawl/crawl/pull/644 * 0.21-a0-405-g9845ef8: Adding more id_floor_items() 10(25 minutes ago, 2 files, 8+ 0-) 13https://github.com/crawl/crawl/commit/9845ef85efa6 07:36:32 |amethyst: would tiles support for non-monospace text be something of interest? 07:40:07 I'm not sure which aspects of the tiles UI are deliberate aesthetic choices, and which are compromises with technical limitations (e.g. the CRT thingy) 08:39:51 !crashlog THUG4L1F3 08:39:52 2. THUG4L1F3, XL2 MiCK, T:544 (milestone): http://crawl.berotato.org/crawl/morgue/THUG4L1F3/crash-THUG4L1F3-20171101-010630.txt 08:40:28 -!- advil_ is now known as advil 08:52:47 Hepl Ally Willingly Steps in Conjured Flame 13https://crawl.develz.org/mantis/view.php?id=11271 by MainiacJoe 09:38:50 THUG4L1F3 (L2 MiCK) (D:1) 09:41:13 !crashlog THUG4L1F3 09:41:14 3. THUG4L1F3, XL2 MiCK, T:544 (milestone): http://crawl.berotato.org/crawl/morgue/THUG4L1F3/crash-THUG4L1F3-20171101-133850.txt 09:47:00 re the question (from a few days ago) of where discussions like https://crawl.develz.org/tavern/viewtopic.php?f=8&t=24461 should happen, I'd actually be in favor of using issues for something like that (if others could be swayed) 09:47:16 I guess traditionally this would happen on the mailing list? but no one seems to use that any more 09:48:41 tbh I think tavern often poisons game design-related discussions, and I personally don't like to participate in real discussions there 09:55:15 advil: well, some of us do like to read and participate on GDD 09:55:48 I can understand how other devs might not want to certainly, and there's no requirement to discuss things there 09:56:13 I think the issue about using "issues" (heh) is that it's not great as a bug tracking tool 09:56:44 Which was more of a technical problem; there wasn't a way to submit bug reports there that didn't involve trying to use some strange issue pre-formatted template, as I recall 09:57:57 The concern about using issues was a duplication between mantis and github, with no good way to cross-reference the two (e.g. marking things as duplicates) 09:58:44 yeah, switching to issues for bug tracking seems unlikely 10:01:08 I don't mean to object to someone posting something like yermak's post there, I just am not personally as likely to respond there, and I guess I haven't really viewed GDD as a dev forum 10:03:49 I'd like to have some place where I could get answered by devs. 10:12:02 this channel is possibly the best place? ...I guess you may want to keep in mind that things have been somewhat quiet in dev-land overall lately 10:12:54 yes, devs will certainly answer if you ask question, so this is a good place to ask 10:13:06 I do actually read gdd, I just don't like to post there, so someone like me might respond here 10:13:52 yeah, not posting a lot is pretty common for most devs I think, but it's probably the best place we have to read discussions between players that are game design related 10:15:15 I've heard reddit brought up as some kind of "alternative" presumably because upvoting is supposed to help, but you don't really read game design discussions there and there's much less a critical mass of discussion related to game design involving really experienced players 10:15:48 So I like to read reddit as well, but I don't find it to be a very good forum for discussing game design 10:16:41 ##crawl can be nice because of long-time players being there, but's irc, not a discussion forum with permanent posts 10:38:25 I have a feature I'd like to work on, just want to see what people here think first - machine parsable morgues that get dumped along with the human readable ones. Reasonable? 10:44:09 sounds like it would double disk usage from morgues, which isn't necessarily prohibitive, but it's a factor 10:44:37 is this something that would introduce a dependency into crawl? not sure what kind of format you'd be using 10:46:55 I was thinking about the disk usage factor - we could make it a flag/option for server ops? As for introducing a dependency/format I'm not sure yet. JSON is one possibility, or we could come up with our own format 10:47:53 On login, gnoll skills at 27 show 3% train until XP is earned 13https://crawl.develz.org/mantis/view.php?id=11272 by Spathi 10:50:00 it's difficult in-part if you don't also introduce some kind of application that uses this at the same time 10:50:44 we've talked about perhaps trying to move more data into the logfile 10:50:54 so that we could do more detailed sequell queries, for instance 10:51:03 and this would also enable tools to be made that would parse this data 10:51:44 I'm not sure if trying to make a second logfile instead (logfiles are deliminated one-line-per-entry files that the Sequell irc bot parses to allow queries against its database) 10:52:46 there are probably some representation challenges with that format, and I'm not sure if Sequell is the best place for all that data 10:53:07 and even then, probably sequell would need some changes to support this data 10:53:54 But I'm not sure how various admins would feel about having an additional file generated for every game if there's no current application using that data 10:54:38 I guess it's not all that different from keeping around .lst files that mostly go unused 10:55:24 This spawned because I have been working on a DCSS scoring site generator, and was getting frustrated parsing the morgue files :) for some background / my use case. Good to know about the logfiles, I'll take a look at that 10:55:59 Would this scoring site generator make pages based on games won etc? 10:56:24 ideally yes! right now it's just a simple hiscore list and some extra pages 10:56:29 Not sure of the scope of your project, but we're interested in replacing the CAO scoring pages with something better written 10:56:34 ??score 10:56:34 scoreboard[1/1]: CAO scoring: http://crawl.akrasiac.org/scoring/overview.html ; New scoreboard: https://scoreboard.crawl.develz.org/ 10:56:44 that first link is what we're currently useing 10:57:14 that second link is/was a project to revamp the system, but development on it sort of came to a halt 10:57:59 both approaches involve parsing the logfile I mentioned (samed one sequell uses) and putting data into a postgres database (I think the new scoreboard uses sqlite3) 10:58:28 On a semi-related note, I was interested in the possibility of including player doll info in morgue files 10:58:50 player doll info? 10:59:11 So that the highscores list could show player dolls, similar to how the main menu shows them for in-progress games 10:59:26 DenialAdams: so I should say that both involve parsing logfiles and putting them into a database. depending on how your project works out and how much effort you're willing to put into it, you might be able to help us revamp scoring 11:00:13 Ah, I haven't seen the second link before! Very cool. That was the sort of thing I had in mind as an end goal. Currently my WIP stuff uses sqlite as well. I would love to help you guys revamp scoring 11:00:17 you can chat with alexjurkiewicz about that "new scoreboard" project if you want to see the approach he was working on (he has a discord for australian/new zealand crawl players where they discussed that project in the past) 11:00:59 yeah, you might at least familiarize yourself with alex's project and codebase; he kind of lost interest working on all the related issues, but that might even be a good starting point for you 11:01:04 for the current CAO scoring code, it's... 11:01:10 ??score 11:01:10 scoreboard[1/1]: CAO scoring: http://crawl.akrasiac.org/scoring/overview.html ; New scoreboard: https://scoreboard.crawl.develz.org/ 11:01:13 ??score[2 11:01:13 score[2/3]: modified XP + runes * (runes + 12) * 1000. For winners, add 254000 + 2000 * runes + 6250000000 * runes^2/turns. 11:01:15 ??score[3 11:01:15 score[3/3]: Modified xp = 0.7 * xp up to 250,000 + 0.4 * xp between 250,000 and 1M + 0.2 * xp between 1M and 3M + 0.1 * xp above 3M. For an XL27 character, this works out to roughly 0.5M + XP/10. Note that there is an XP cap of 9M (but not in 0.16). 11:01:39 https://github.com/crawl/scoring 11:01:41 there 11:01:53 that's a repo with the current python code to generate pages 11:02:15 and finally here is some similar code we use to score tournaments: https://github.com/crawl/dcss_tourney 11:02:41 which involves a very similar approach (logfile/milestones into database, code to generate html pages from templates) 11:02:55 the tournament website is not operational at the moment, so I can't show you what the tourney pages look like 11:03:00 but they're pretty nice 11:04:38 Thanks for all of the references! I also found a link to the new scoreboard repo (https://github.com/zxc23/dcss-scoreboard). My current WIP stuff was written in Rust - there is one process that watches for new morgue files, parses some limited info out of them and stores it into DB, then another process that serves templated HTML files with info from the DB 11:04:52 Where can I find the log files you're referencing? :) 11:04:58 ??logfile 11:04:58 I don't have a page labeled logfile in my learndb. 11:05:05 hrm, is there a good reference on that 11:05:12 let me give you the sequell config yml file 11:05:23 thank you! 11:05:29 here is the repo for sequell: https://github.com/crawl/sequell 11:05:37 https://github.com/crawl/sequell/blob/master/config/sources.yml 11:05:46 is probably the best place to find each server's logfile 11:05:52 <|amethyst> https://www.reddit.com/r/dcss/comments/5razuf/analyzing_species_in_dcss_over_a_million_games/dd60yem/ 11:06:23 <|amethyst> more information there, including a link to (some) logfiles/milestones that can't be downloaded anymore 11:06:28 http://crawl.berotato.org/crawl/meta/git/ is an example "meta" dir with the logfile and milestone file (and other files for e.g. sprint) 11:06:35 oh, yeah 11:06:44 <|amethyst> I still haven't gotten around to bugging my ISP about CSZO, sorry :( 11:06:59 no worries! we all know that I won the 0.20 tournament anyhow 11:07:25 amalloy came in 27th, it was a pitched battle for the ages 11:07:37 pitched? 11:07:48 pitched. 11:08:49 Ahh, so the logfiles are not per player? Also, thanks |amethyst for the reddit link. I'm happy to know I'm not the only one who was looking for this 11:09:02 right, logfiles are for *every* game in that version of crawl (so one for trunk, one for 0.20, etc) 11:09:09 on game is one line in the file 11:09:29 but it only contains a specific set of data, not the full data seen in morgues 11:09:34 !lg . won 11:09:36 74. gammafunk the Phalangite (L27 MfAM of Trog), escaped with the Orb and 3 runes on 2017-06-11 12:18:36, with 1538107 points after 84750 turns and 5:42:55. 11:09:45 sequell parses these and stores the fields in its db 11:10:04 !lg . won x=ac,mhp,score,start,ev,fifteenskills 11:10:05 74. [ac=42;mhp=246;sc=1538107;start=2017-06-11 04:04:09 [20170511040409S];ev=19;fifteenskills=Fighting,Polearms,Armour,Dodging,Evocations] gammafunk the Phalangite (L27 MfAM of Trog), escaped with the Orb and 3 runes on 2017-06-11 12:18:36, with 1538107 points after 84750 turns and 5:42:55. 11:10:25 those data in brackes are some of the data the logfile knows about 11:10:54 but I can't query for e.g. my wielded weapon or my inventory, or even my skills on an individual basis 11:11:07 I just have the 'fifteenskills' field that's a list of fields at level 15 or higher 11:11:22 the cool thing about his being I can do sequellese queries against this db 11:11:35 !lg . won x=avg(mhp) 11:11:36 74 games for gammafunk (won): avg(mhp)=220.82 11:11:43 that's the the average max hp of my won games 11:12:25 but for scoring pages we're just trying to get a tabulation of the players games and wins, along with some basic stats like winrate, total score, etc 11:13:07 since sequell is maintained and operated by the legendary green.snark, we more or less regard it as the 'official' score tracking system 11:13:23 since it doesn't go down, is well maintained, and has a lot more functionality 11:14:03 it does go down somtimes and there can be maintenance issues, but the CAO scoring is markedly worse in those regards 11:14:38 oh, very neat! So the question is if we want to add more information to the log file (that may or may not "just work" with sequell and the existing stuff) or add some machine readable morgues and revamp the CAO scoring? 11:14:41 mostly due to its implementation not being very robust and there being no one who has the time to work on it 11:15:14 well, I would say that revamping scoring is basically an independent issue; it's something we'd definitely greatly appreciate and is more important than adding morgue parsing 11:15:48 ok, understood. Because the scoring can also operate off of the log files? 11:15:50 for morgue parsing, I'm not sure if sequell is a great candidate for that data since sequell has to be queried and already can be pretty slow for a lot of queries 11:15:55 yes, definitely 11:16:03 I would say that it should operate off the log files 11:16:26 though I suppose if you added this morgue representation it could operate off of that 11:17:22 if you wanted to have more detailed kind of stats based off data in morgues that's not in the logfile, then it might be a pain working with both 11:18:08 regardless, if you could get scoring pages for us that were reliable and correct 11:18:14 having all the basic details tracked by CAO scoring 11:18:23 that would be greatly appreciate by devs and players alike :) 11:18:26 *appreciated 11:19:17 Artificer/Demonspawn "accessibility" 13https://crawl.develz.org/mantis/view.php?id=11273 by FerrousOxx 11:19:23 http://crawl.akrasiac.org/scoring/index.html 11:19:30 is all the other kinds of data CAO tracks 11:19:43 Ok, makes sense :) I don't mind using the log files necessarily, it just wasn't what my first thought was. It sounds like getting better scoring with the information available right now would be where to start 11:19:57 yeah, that would be wonderful 11:21:40 Can you provide me with some sort of list of what would be desired traits in the new scoring system that are lacking now? 11:23:30 |amethyst might be better able to comment on some of the technical issues, and I think the main problems with CAO scoring are related to that 11:23:56 it seems to frequently go down, and if there's a problem with the database, rebuilding it takes ages 11:24:30 ok, thanks! performance seems like a factor to keep in mind then. 11:24:44 DenialAdams: you can compare the new scoreboard page with the current page to see some improvements that they developed 11:25:02 https://scoreboard.crawl.develz.org/players/gammafunk.html 11:25:18 a better overall display of the data 11:25:30 comapred to here: http://crawl.akrasiac.org/scoring/players/gammafunk.html 11:25:41 also note that CAO reports 83 wins, which is incorrect 11:25:53 some wins got duplicated in the CJR server logfile 11:26:10 those were since removed, but we haven't rebuilt the CAO database since that correction, because rebuilding takes a long time 11:26:42 sequell has a correction where if it sees a duplicate logfile entry, it simply doesn't added (this was added fairly recently to sequell due to this problem on CJR) 11:27:08 s/doesn't added/doesn't add the entry/ 11:27:19 by that I mean it doesn't add the duplicate 11:27:44 that's something that would be nice for this scoreboard 11:28:09 some functionality we'd like to have: something like sequell nicks for the scoring pages 11:28:13 !nick gammafunk 11:28:13 Mapping gammafunk => (!gfspeed) gammafunk 11:28:29 I have done lots of speedrun games, which I don't like to include in the stats for my casual games 11:28:49 hence I have a sequell keyword 'gfspeed' to exclude those games 11:28:53 !won gammafunk 11:28:59 gammafunk has won 74 times in 776 games (9.54%): 2xOpTm 2xSpEn 1xBaNe 1xBaWn 1xCeAM 1xCeAr 1xDDGl 1xDDHe 1xDEAs 1xDEEn 1xDgEE 1xDgWn 1xDrCK 1xDrMo 1xDrNe 1xDrTm 1xDsDK 1xDsFi 1xDsNe 1xFeCK 1xFeCj 1xFoAK 1xFoVM 1xGhEE 1xGhMo 1xGrAs 1xGrDK 1xGrIE 1xGrWn 1xHECK 1xHEFi 1xHEWr 1xHOAs 1xHOFE 1xHOWz 1xHaAM 1xHaAr 1xHaCj 1xHuNe 1xHuSk 1xHuSu 1xHuTm 1xKoAr 1xKoEn 1xKoHu 1xKoMo 1xMfAM 1xMfSk 1xMfWz 1xMuGl 1... 11:29:00 !won :gammafunk 11:29:01 :gammafunk has won 77 times in 3924 games (1.96%): 2xOpTm 2xSpEn 1xBaNe 1xBaWn 1xCeAM 1xCeAr 1xDDFi 1xDDGl 1xDDHe 1xDEAs 1xDEEn 1xDgEE 1xDgWn 1xDrCK 1xDrMo 1xDrNe 1xDrTm 1xDsDK 1xDsFi 1xDsNe 1xFeCK 1xFeCj 1xFoAK 1xFoVM 1xGhEE 1xGhMo 1xGrAs 1xGrDK 1xGrIE 1xGrWn 1xHECK 1xHEFi 1xHEIE 1xHEWr 1xHOAs 1xHOFE 1xHOWz 1xHaAM 1xHaAr 1xHaCj 1xHuNe 1xHuSk 1xHuSu 1xHuTm 1xKoAr 1xKoEn 1xKoHu 1xKoMo 1xMfAM 1xMfSk... 11:29:16 that second one ignores my nick and uses just my player name, so it's including those games 11:29:20 ah, I see. Where do you put the gfspeed keyword? 11:29:44 this is pretty advanced functionality and I'm not sure it's possible to put things like that in your scoreboard 11:29:49 !kw gfspeed 11:29:49 Keyword: gfspeed => char=ddfi|deie|desu|grgl|heie|hesu|hogl|mffi|mfgl|mfbe|mibe|migl|musu|miwr|trck|vsie|vsfi|vssu !gfslow 11:29:59 you can define keywords in sequell for any kind of query 11:30:01 ??userdef 11:30:01 userdef[1/1]: https://loom.shalott.org/userdef.html 11:30:16 that's the current list of all sequell commands, functions, and keywords 11:30:26 that are user-defined (like gfspeed) or built-in 11:30:40 we've talked about actually moving page generation into sequell itself 11:30:49 in that this would be idea in a lot of ways 11:31:18 sequell is very advanced and supports making arbitrary functions to query the data 11:31:46 in a perfect world we'd be able to have an html page for my gammafunk account that respects this nick 11:33:18 I see. So unless the scoring pages are built into sequell, there would need to be some duplication of sequell features 11:34:08 Yes, certainly some duplication, but to be clear you don't have to add sequell-like functionality to scoring to make the project useful to us, certainly not at first, at least 11:34:58 Having more reliable scoring pages along would be very helpful, and any functionality that could be added on top is just a bonus 11:35:03 s/along/alone/ 11:38:22 Ok :) I think I may have gotten in over my head, but I really would love to tackle this. I'm going to take a look at the new scoring, old scoring, and sequell and see how I want to try and improve things 11:39:05 thanks a lot, DenialAdams 11:40:26 no problem! Thank you for all of the great info. I'm sure I will be back soon with questions 11:51:08 I already have one question - sequell and the scoring both maintain their own database. Could they share one? 11:52:17 I feel like ideally creating the database and reading it (for scoring or sequell) could be two seperate concerns 11:54:09 But I suppose if scoring and sequell run on two different servers that could be an issue 11:55:04 even in that case having a common library/binary to build the database seems like an improvement, sharing code 12:17:48 DenialAdams: well, that was why I mentioned at one point the idea of having "sequell" generating the scoring pages 12:17:58 but yeah, you could potentially use the same database 12:18:12 this might allow you to implement some of that nick functionality I mentioned for the pages 12:18:40 but on a more practical level, sequell is maintained *and* hosted by greensnark 12:19:26 if you actually just share the code and run the scoring pages based on that shared code, that could be fine as well 12:20:12 I'm not sure if trying to host scoring pages on shallot.org is either a good idea or one that darshan (greensnark) would agree to 12:21:01 I think that sharing code would in general be good though, yes, as long as you find sequell's code usable 12:38:01 gammafunk: thanks again for your insight. I think I'll start by generating my own database, try to make it fast, and see if I can share that later 13:04:46 Unstable branch on crawl.akrasiac.org updated to: 0.21-a0-405-gf2af9a1 (34) 13:24:12 -!- amalloy_ is now known as amalloy 14:15:21 rsync to new server has started. We may be looking at some data loss and me having to set up crawl from the ground up minus some config. 14:15:23 Welp. 14:25:55 TZer0: any word as to what kind of data loss? 14:27:37 It seems like one drive in my raid has just completely failed. 14:27:37 I'm not 100% certain about this 14:27:43 And there may be some data loss of random files on the other one 14:27:50 I'll get back to this once the rsync is done 16:28:15 03gammafunk02 07* 0.21-a0-406-gc134197: Add a -force-map option for objstat and mapstat 10(2 weeks ago, 6 files, 60+ 10-) 13https://github.com/crawl/crawl/commit/c13419712335 16:28:15 03gammafunk02 07* 0.21-a0-407-gc23e797: Give air elementals Vortex, a weakened form of Tornado 10(3 weeks ago, 10 files, 111+ 34-) 13https://github.com/crawl/crawl/commit/c23e79788af5 16:28:15 03gammafunk02 07* 0.21-a0-408-gbd10913: Show wind immunity in monster info 10(3 weeks ago, 1 file, 4+ 1-) 13https://github.com/crawl/crawl/commit/bd1091303caf 16:28:15 03gammafunk02 07* 0.21-a0-409-g7d0f330: Give Sojobo wind immunity 10(3 weeks ago, 1 file, 3+ 2-) 13https://github.com/crawl/crawl/commit/7d0f330acb1c 16:28:15 03gammafunk02 07* 0.21-a0-410-g2694e31: Revamp of the Cloud Mage WizLab 10(2 weeks ago, 1 file, 81+ 69-) 13https://github.com/crawl/crawl/commit/2694e31eb0dd 16:28:15 03gammafunk02 07* 0.21-a0-411-gdc66bd4: Give the Cerebov statue tile a DES name 10(3 hours ago, 1 file, 1+ 0-) 13https://github.com/crawl/crawl/commit/dc66bd4d4683 16:28:15 03gammafunk02 07* 0.21-a0-412-ga58e5a4: Adjust the Hellbinder's spell spell set. 10(3 hours ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/a58e5a46b323 16:28:15 03gammafunk02 07* 0.21-a0-413-g4a39e9c: Rework of the Hall of the Hellbinder 10(7 days ago, 1 file, 87+ 88-) 13https://github.com/crawl/crawl/commit/4a39e9cdf1c3 17:17:18 The build was fixed. (master - 4a39e9c #8880 : gammafunk): https://travis-ci.org/crawl/crawl/builds/295938214 17:57:57 <|amethyst> Teleport (or in this case, shaft) closet: http://crawl.akrasiac.org/saves/Neil-crawl-git-f2af9a17fc-171101-1757.tar.bz2 17:58:06 <|amethyst> fortunately I have ?tele 19:04:40 03|amethyst02 07* 0.21-a0-414-gba8bdf1: Fix teleport closets in water_boxes_lemuel. 10(2 minutes ago, 1 file, 1+ 0-) 13https://github.com/crawl/crawl/commit/ba8bdf1b2e0d 19:08:18 <|amethyst> IMO that vault's kind of silly and should probably go 19:08:40 <|amethyst> it takes up a bunch of space, and its purpose seems to be "free items to those with flight/swimming" 19:09:10 <|amethyst> OTOH it doesn't have no_monster_gen on the water, so at least it can spawn eels 19:09:39 <|amethyst> maybe replace the fungi with oklobs? 19:11:04 <|amethyst> or burning bushes, since they'll make LOS-blocking steam and annoy ranged/spell users 21:18:08 -!- amalloy is now known as amalloy_ 21:54:18 som it;s bveing reported over in ##crawl that footv is stuck. does varmin need to be restarted or ?? 21:54:24 *so, it's 22:10:43 !lg * -tv:nuke 22:10:45 FooTV playlist clear requested by *. 22:22:39 yeah it has been stuck since yesterday