00:03:22 i think a more prompt change would be relatively low-risk; the relevant code in messages seems reasonably straight-forward 00:04:13 I'm talking about testing by users :-) 00:04:27 the risk is that an alternative wouldn't work very well 00:04:59 and/or be even worse for newer players 00:05:25 hm, it's possible flashing the prompt wouldn't work, but I think some form of attention-getting visual feedback is required, and I don't see how to do it any other way 00:06:26 flashing focus-grabbing ui elements is also a common pattern at the OS level, e.g. for dialog boxes 01:38:47 Unstable branch on crawl.develz.org updated to: 0.25-a0-930-g9f5dfa9815 (34) 01:45:00 New branch created: pull/1384 (1 commit) 13https://github.com/crawl/crawl/pull/1384 01:45:00 03Quipyowert202 07https://github.com/crawl/crawl/pull/1384 * 0.25-a0-870-ge66962a: Fix resource leak if allocating PNG structs fails. 10(27 minutes ago, 1 file, 4+ 0-) 13https://github.com/crawl/crawl/commit/e66962abe17a 02:12:10 Windows builds of master branch on crawl.develz.org updated to: 0.25-a0-930-g9f5dfa9815 03:28:17 -!- The topic of ##crawl-dev is: Crawl Development | Logs: http://s-z.org/crawl-dev/ | People with +v have commit access. | Please keep general Crawl-related chat to ##crawl. | Dev wiki: http://crawl.develz.org/wiki | Long stuff to a pastebin service, please. 03:28:31 -!- The topic of ##crawl is: Play Dungeon Crawl Stone Soup online now! Type ??servers for instructions. | http://crawl.develz.org | FooTV: http://termcast.develz.org - ??footv for instructions | See also ##crawl-offtopic 03:41:48 Fork (bcadrencrawl) on crawl.kelbi.org updated to: 0.22.1-2664-g56b5171bbf 05:28:41 should jorgrun still spawn in swamp given LRD is pretty bad there? 05:28:55 also man, that's a lot of spells 05:29:10 does he really need grasping roots AND petrify? 08:01:12 FR: Qaz storm noise attract attention towards storms and distract from the player. :) 08:05:09 Like the storms also somewhat hide the player amid the noise and obfuscation from the storms. 08:07:53 Would give Qaz an extra dimension of playability. 10:39:09 I know I say this every once in a while but saving every ttyrec as a policy is insane 10:40:04 also ttyrec is a terrible format 10:40:15 (ok I'm done for now) 10:41:05 moving on from my venting, sequell is really having issues 10:41:25 I've been getting lag on very normal commands, e.g. !source 10:42:17 just got a timeout on `!lg * boring` 11:18:21 so one obvious policy tweak that might even be implementable on the webtiles server itself would be to not save ttyrecs for `boring` games 11:18:51 I think sequall might just be dead right now 11:22:45 I've fleshed out a bit more detail on a proposed new tournament scoring 11:22:45 https://github.com/crawl/crawl/wiki/New-Tournament-Scoring 11:42:15 Just to keep the ball rolling, it would be good to get some initial team feedback if this is the right direction and shape this into a form suitable for player feedback. 11:43:55 this looks like a good thing to me 11:44:32 one thing to check is what this ends up looking like for players not in the top 100 or whatever 11:44:49 also for players who don't win 11:48:09 a very rough estimate from t0.24 is that something like 50% of players who have score > 100 never won, but still have score from runes, ghosts, branches, uniques, etc 11:50:53 probably some huge proportion of players have <5 wins 11:51:03 it's annoying to try to calculate without sequell 11:53:28 so one advantage of the current scoring system, I think, is that though the vast majority of players have no chance of getting in the top 100 or whatever, they still have a fun mini-game of watching themselves in the leaderboard (and from talking to people last time around I think this actually *is* fun, I wasn't initially sure) 11:53:50 the new approach would get this to some degree but I don't yet have a sense for how much 11:54:28 and from looking at the current category list I would guess that non-winners are a bit downweighted relative to the current system? 12:03:33 Yes, non-winners are a bit downweighted 12:03:46 but they'll all be tied in the categories requiring wins 12:04:03 and move up and down based on combo high scores (which don't have to be wins) and banner points 12:04:25 the draft list of banner points is mostly all bounded, but a variety of things to try to do 12:04:59 the Ash banner could be changed to be closer to the old rune+branch scoring to give a bit more movement 12:11:50 yeah ashenzari does look like a good banner for my aims 12:12:09 and yred 12:15:00 not about that issue, but any thoughts on having a banner that is scored by win ratio? 12:15:23 re nemelex choice, I do think wherever possible we should keep the soft caps 12:19:16 Could be fun? I worry about grinding it to keep it maximised, but I think doing that requires giving up on some of the higher risk point sources 12:28:54 I hesitate making the Ash banner unlimited, since it will come to eclipse the other banner scores with sufficient grinding 12:29:00 and a logarithm just increases the grind threshold 12:39:01 yeah, unlimited seems too much 12:39:30 but I think adding branch entry/end to Ash's banner could be nice for players in this category 12:40:06 I could be wrong but I don't think grinding would really help with a win ratio banner past a certain point, and it would reward people who do well at a small number of games 12:40:13 full disclosure I have a coi with this suggestion :-P 12:41:06 I wonder if having the ash banner as a category would be good? 12:41:36 actually since some of your other categories line up with banners that have no points why not just directly call them banners? 12:42:09 I guess it makes it weirder to have some banners collapsed and some not 12:43:59 one other reason I was thinking about a winrate banner is because if you look at top players from prev tournaments there's very interesting differentiation there that doesn't line up with overall ranking (though of course they weren't targeting winrate so their behavior could well change) 12:45:57 yeah 12:56:00 flavor wise, perhaps jiyva could be the one awarding distinct streaks and flexibility and chei the winrate parameter 12:56:42 yeah I was thinking chei also for that 12:56:50 I'm fine if you don't think it's a good idea to be clear :-) 12:57:18 another big category of player distinct from non-winners is people who don't have a lot of time to play 12:57:36 No, you've convinced me because of that player category 12:57:51 ah ok good 12:58:00 I wonder if there are other things to do for that category? 12:58:16 though we may get higher engagement on a pandemic tournament 13:04:38 Digory (L18 MiBe) ASSERT(!invalid_monster(&mons)) in 'mon-death.cc' at line 1710 failed. (Elf:3) 13:32:26 who picked this naming scheme for ttyrecs 13:48:05 t e c h d e b t 13:49:50 I know I'm usually the "oh we must test this endlessly" person but for ttyrecs I honestly just want to break a bunch of stuff and let someone else sort out sequell, tv, etc 13:49:50 (I'm not going to, but...) 13:57:36 how important are ts files again? 14:51:08 03ebering02 07* 0.25-a0-931-g5a33db6: Remove targeting from Summon Lightning Spire 10(62 minutes ago, 4 files, 10+ 45-) 13https://github.com/crawl/crawl/commit/5a33db65c3ef 14:51:08 03ebering02 07* 0.25-a0-932-g27cb841: Prevent Eldritch Tentacles from attacking out of LOS 10(20 minutes ago, 1 file, 3+ 1-) 13https://github.com/crawl/crawl/commit/27cb841e2324 15:06:04 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-932-g27cb841e23 (34) 15:19:38 re sls, is it ok that you can just keep recasting to reroll the position (each new cast dispels the old one)? costs mp so that be sufficient 15:19:59 yeah I think that's fine 15:22:00 *so that may be 15:22:58 I guess you can also just hit it to force-dispel it if you get a bad positioning and are out of mp 15:25:17 I didn't want to be sso severe as to force a player to live with the cast position they get 15:25:27 yeah, seems fair 15:25:29 just make getting the best position a bit harder than picking it 15:25:54 my initial question was can you use it in a hallway and it seems like there are many ways to try 15:26:22 I wonder if given that dispelling a summon is kind of trivial now if you are next to it, it would make sense to provide a more obvious way to do it ui-wise? 15:42:55 just make getting the best position a bit harder than picking it 15:43:01 oops 15:43:21 gammafunk: not sure how often or thoroughly you read scrollback but I've elaborated some on possible new t rules 15:43:46 cool, will take a look after I finishling wrangling this cerebot code 16:23:40 Escape doesn't return to the ctrl-f results anymore 13https://crawl.develz.org/mantis/view.php?id=12250 by nstoddard 17:09:54 aidanh you're the primary dev behind UI right? I'm having a weird bug I need help with/can't explain that's webtiles UI specific. stuff that's showing up fine on local coming up as ��4?>V on webtiles and similar. 17:11:11 Which is reproduced in vanilla by doing... 17:13:08 I have no idea. I'm legit asking for help with my fork at this point. >_< 17:16:30 https://hastebin.com/rakokobadu.cpp 17:17:47 `return ostr.str().c_str()` 17:17:58 see first answer here: https://stackoverflow.com/questions/22330250/how-to-return-a-stdstring-c-str 17:18:35 basically you are returning a pointer to a deleted/freed object and so it's printing random garbage from memory 17:19:16 03ebering02 07* 0.25-a0-933-gf879fa1: Improve foxfire and ball lightning target acquisition (Lici the Crawler) 10(19 minutes ago, 2 files, 12+ 7-) 13https://github.com/crawl/crawl/commit/f879fa1a088c 17:19:16 03ebering02 07* 0.25-a0-934-g041635d: Reduce the delay before an eldritch tentacle enters 10(5 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/041635d5d19f 17:19:37 the difference in builds is accidental, and something extra is presumably happening in the newly free memory in webtiles, but it's not safe to return c_str() either way 17:21:24 I'd recoded it a few times trying to fix it without knowing what was wrong. A previous version was makestringf() instead of ostr...same issue I suppose. 17:28:42 kate-: do you have thoughts on tournament rules (sorry to be ping happy on this but there is a bit of a timeline on this and I don't want to repeat the past two tournaments where we ran out of inertia) 17:30:48 btw for anyone wondering what the notice on cao is about, I've started (slowly) uploading some select <2018 ttyrecs from cao into a b2 bucket; eventually I'll add the new location to sequell but not yet 17:31:19 I'd like to get everything <2018 off the server but it's going to be a s l o w process 17:31:45 I'm starting by targeting the worst accounts 17:33:11 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-934-g041635d5d1 (34) 17:34:35 ebering: hmm, i've never really followed tourney rules/scoring that closely so i don't have very strong opinions. i do like the idea that was discussed the other day of some kind scoring based on rank in multiple categories 17:36:10 but i don't think i'd be much help in terms of specifics i'm afraid, it's been a good few years since i've even played much in a tourney 17:36:51 thanks! 18:04:26 advil one of the places this bug happens can't be explained that way. https://pastebin.com/raw/XP6XsqjG 18:15:33 returning a const char * from _seekspell is fine if the struct was initialized correctly, so that code snippet is probably not the locus of the bug 18:16:05 returning c_str isn't bad because it's a const char *, it's bad because it's a const char * to something that will get deallocated on the return 18:20:55 what I would check is how chaosTitle is getting initialized 18:21:14 my guess is you are doing something similar, e.g. setting it to the return value of c_str() for something that is getting deleted 18:23:26 advil: the ts files are necessary for syncing playback of ttyrec in footv; I don't know fully how, but I know that those files being missing on cwz (because they're in the wrong dir) means that most ttyrecs don't play properly 18:23:39 sad 18:23:42 Unstable branch on underhound.eu updated to: 0.25-a0-934-g041635d5d1 (34) 18:24:05 you might need to dig into that yourself if you want an actually informed answer, since I don't have one 18:24:20 I know those files are fairly small though, compared to ttyrecs 18:24:33 yeah, size isn't really the issue, but magnitude of the number of files 18:25:08 and figuring out which files might need to be moved together 18:25:51 I think FOOTV only knows to look in the same dir of the ttyrec 18:26:11 life would be somewhat better if all the replay-related files for each game could be bundled into a .tar or something like that (with a sane filename) 18:26:21 https://github.com/greensnark/dcss_tv snarks repo 18:26:35 I say somewhat because there'd still be an absurd amount of them 18:27:58 are you following the tiles replay thing? 18:28:34 https://www.reddit.com/r/dcss/comments/gbek1p/dcssreplay_07_beta_release_cached_edition/ 18:29:11 I guess it's windows-only 18:31:51 I am in that I got it to start in wine and render the opening frame of the ttyrec 18:31:59 sadly playback doesn't work properly from there 18:32:16 the author said he's probably going to make it web-based to get around combatibility issues 18:32:26 yeah I saw an issue for that 18:32:39 it looks like it'd need a bit of a rewrite for that, though 18:32:51 it's a cute approach, I guess, but seems pretty fundamentally limited 18:33:10 how so? 18:33:21 I mean, it'll never be able to reconstruct a real tile view of course 18:33:29 I mean, just accurately rendering tile based on a ttyrec glyphs, yeah 18:33:34 *tiles 18:34:15 I assume we'd ideally have some kind of playback client that just read json and had cached clients for playback from the various versions, I guess 18:34:37 sounds utopian 18:35:56 I guess for just map replays in a way it wouldn't really be too complicated what needs to be stored 18:36:10 abut tournament scoring, it might be a good idea to restrict players to a single server 18:37:14 yeah, storage of json sounds better than storage of ttyrecs, but even I can see that there are issues with having a client that plays back the json 18:37:28 even I, totally ignorant UI person that I am 18:37:52 for example, if one scoring category is "time of first win", it would make sense to play a game to be standing on the d:1 doors, then start a new game elsewhere and try to do it faster. Decide which game to finish first based on their time elapsed 18:38:15 (assuming time of first win is duration of the game, not "finished earlier in the tournament") 18:38:36 time of first win is order of finishing in the tournament 18:38:44 the JSON data is significantly larger than TTYREC data, FWIW 18:39:04 yes, the json data as used over websockets wouldn't cut it 18:40:06 I don't think we want to limit play in tournament to one server due to the hassle that entails for players, but even putting that aside, that category isn't susceptible to cheating by using multiple servers 18:41:23 fwiw, i've been tossing up disabling ttyrecs on cpo completely. If people want to record a game, they can use streaming services. Not many games support replay systems, and I think the value is limited when ttyrecs record in a way that a decreasing percentage of the playerbase understands 18:41:23 the only categories that were negatively influenced by multiple servers were the god/species/class points where playing earlier meant more points 18:41:23 and from what I understand, those points would go away in the new format 18:41:32 if there was a "tiles compatible" replay system I'd definitely enable that, but I think the days of unlimited retention are gone for CPO 18:41:54 they were criticized a lot as it is, not just because they encouraged parking of games but also because it felt kind of arbitrary 18:43:53 advil: well, I guess I'd be surprised if at least a subset of json wouldn't work for playback and be less when compressed than ttyrec, but maybe it's utopian dream like you said 18:47:12 map json is a bit massive, my guess is that it wouldn't be smaller when compressed 18:48:04 smaller than ttryecs that is 18:48:20 rewrite webtiles communication in protobuf!!!11 18:49:54 bh would be so happy if we did 19:11:33 advil they are exactly the same as the normal titles. https://pastebin.com/fqwzQmrc 19:16:40 looking at all the FooTV perl code snark wrote over the years to do all sorts of ttyrec frame mangling I'm both amazed and horrified 19:16:43 mostly amazed 22:03:18 03advil02 07* 0.25-a0-935-ga522a9f: Fix description of what happens when attacking summons 10(6 minutes ago, 1 file, 23+ 14-) 13https://github.com/crawl/crawl/commit/a522a9f31ccb 22:18:46 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-935-ga522a9f31c (34) 22:20:30 03advil02 07* 0.25-a0-936-gcfa9253: Fix a pronoun case 10(75 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/cfa92534785f 22:33:16 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-936-gcfa9253478 (34) 22:36:21 03gammafunk02 07* 0.25-a0-937-gcfae464: Remove special flag from the Sceptre of Torment 10(4 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/cfae464cf3d8 22:48:01 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-937-gcfae464cf3 (34)