01:41:28 -!- eb_ is now known as eb 02:45:32 Monster database of master branch on crawl.develz.org updated to: 0.22-a0-402-ga933591 06:10:55 -!- johannm is now known as Tungsten_ 09:11:27 -!- us17_ is now known as us17 09:11:46 -!- TooLmaN_ is now known as TooLmaN 09:12:09 hm 09:12:15 I have split fro myself? 09:12:18 *from 09:12:59 freenode seems toubled 09:13:09 -!- advil_ is now known as advil 09:15:33 does anyone know how the gnoll viable species list came about? 09:15:36 it sems kind of odd 09:15:51 also, I'm not enthralled with the word "viable" there 09:17:17 it is, sk, tm, wr, am, wn 09:17:27 !lg * recent gn-- s=class o=% / won 09:17:28 786/78884 games for * (recent gn--): 13/354x Air Elementalist [3.67%], 35/1142x Venom Mage [3.06%], 20/772x Earth Elementalist [2.59%], 27/1109x Summoner [2.43%], 49/2085x Wizard [2.35%], 14/596x Ice Elementalist [2.35%], 22/948x Berserker [2.32%], 31/1451x Necromancer [2.14%], 31/1515x Arcane Marksman [2.05%], 19/933x Fire Elementalist [2.04%], 13/645x Conjurer [2.02%], 19/952x Chaos Knight [2.00... 09:37:35 03advil02 07https://github.com/crawl/crawl/pull/721 * 0.22-a0-288-gbf7c31b: Change species generation to python 2.7 10(23 hours ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/bf7c31b779ee 09:37:35 03advil02 07https://github.com/crawl/crawl/pull/721 * 0.22-a0-289-g9201aa5: Format/refactor for 80 columns 10(23 hours ago, 1 file, 79+ 40-) 13https://github.com/crawl/crawl/commit/9201aa59e481 09:37:35 03advil02 07https://github.com/crawl/crawl/pull/721 * 0.22-a0-290-g9ef5112: Don't inherit from dict, use constructor instead of factory function 10(21 hours ago, 1 file, 56+ 40-) 13https://github.com/crawl/crawl/commit/9ef5112fbdfa 09:37:35 03advil02 07https://github.com/crawl/crawl/pull/721 * 0.22-a0-291-ge24a7fb: Make build dependencies more specific 10(21 hours ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/e24a7fb3aca0 09:37:35 03advil02 07https://github.com/crawl/crawl/pull/721 * 0.22-a0-292-geca3f6f: Handle empty mutation lists a bit better 10(21 hours ago, 2 files, 17+ 15-) 13https://github.com/crawl/crawl/commit/eca3f6f3564f 09:37:35 03advil02 07https://github.com/crawl/crawl/pull/721 * 0.22-a0-293-g08ee2c8: Add entry format description to header for species-data.h 10(21 hours ago, 1 file, 19+ 0-) 13https://github.com/crawl/crawl/commit/08ee2c857f5f 09:37:35 03advil02 07https://github.com/crawl/crawl/pull/721 * 0.22-a0-294-g648d254: Don't set recommended weapons for non-starting species 10(20 hours ago, 1 file, 16+ 14-) 13https://github.com/crawl/crawl/commit/648d2540c07b 09:37:35 03advil02 07https://github.com/crawl/crawl/pull/721 * 0.22-a0-295-gd0c7f61: Fix undead type 10(20 hours ago, 3 files, 3+ 1-) 13https://github.com/crawl/crawl/commit/d0c7f61ef533 09:37:35 03advil02 07https://github.com/crawl/crawl/pull/721 * 0.22-a0-296-g2525d87: Add walk/pray verbs to SP_UNKNOWN and deprecated species 10(20 hours ago, 2 files, 10+ 0-) 13https://github.com/crawl/crawl/commit/2525d8704819 09:37:35 03advil02 07https://github.com/crawl/crawl/pull/721 * 0.22-a0-297-gd4cb019: Handle demigod stats properly 10(19 hours ago, 3 files, 15+ 6-) 13https://github.com/crawl/crawl/commit/d4cb019f9d34 09:37:35 ... and 2 more commits 10:23:07 -!- TAS-2012v is now known as TAS_2012v 11:27:54 03Aidan Holm02 07* 0.22-a0-403-gc217a56: Fix loading of paletted images on SDL tiles 10(8 days ago, 1 file, 2+ 1-) 13https://github.com/crawl/crawl/commit/c217a56188f2 11:46:49 ??plan 11:46:49 plan[1/2]: https://github.com/crawl/crawl/wiki/0.21-Development-and-Release-Plan 12:10:00 Unstable branch on crawl.jorgrun.rocks updated to: 0.22-a0-403-gc217a56188 (34) 13:05:13 Unstable branch on crawl.akrasiac.org updated to: 0.22-a0-403-gc217a56 (34) 15:14:38 maybe gn should just have all backgrounds be viable, would be thematic 15:23:22 advil: I don't think that's a great precedent since you could use flavor justifications to do various other recommendations for species that way 15:23:41 and the idea was to keep the number of recommendations between species/classes more balanced 15:24:46 I don't know how many species recommend wizard and if there's another background that gnoll recommends that already has a lot of recommendations 15:25:05 so that you could just swap things around a bit 15:33:58 well basically none of the recs make sense, they're all from bultungins (or at least that's what amalloy said on reddit) 15:34:37 well, they don't *not* make sense but they're definitely not the most viable ones for gn 15:35:18 also, I doubt that the existing recs for gn really factor into the balancing across species. (I haven't checked the inverse recs) 15:37:42 the recommendations were never supposed to be the most viable in the sense of the best for winning 15:38:39 all I mean by pointing out the balance of recommendations is that each class recommendation influences the number of species that recommend that background 15:38:59 it is true that changing one doesn't have a big impact, but it may be that e.g. wizard is recommended a lot already 15:39:43 but we obviously don't want to recommend only the most viable things since that will by definition make it impossible to balance those numbers 15:40:13 <|amethys1> gammafunk: but that was a big part of the reason we moved to asymmetric recommendations 15:40:13 sorry? 15:40:34 <|amethys1> so that we didn't have to worry so much about having class X be recommended by only so many species 15:41:25 well, I think we wanted the flexibility to adjust those recommendations so that they were balanced is what I'm saying 15:41:48 if by only so many species you mean too few species recommend a background, then yes I'd agree that's part of balancing those numbers 15:42:42 but just to be clear, I think the recommendation is about recommending things that are interesting and have good synergy with a species 15:43:00 these will often be strong classes but very often they will not be the strongest (hence not the most viable) 15:44:28 right now Gn recommends: Wn Sk Tm Wr AM 15:44:43 Wz has six recommendations 15:45:32 and it looks like most mages have 6, a couple have 5 or 7 15:47:59 so yeah we could simply add Wz to Gn since it only has 5 15:48:19 and it does not recommend any other mage start 15:50:00 Adding wz to gn would make sense, but I'm not sure any of the existing ones make sense, like I said they were apparently constructed for gnolls v.1 15:50:13 also, "viable" is a really bad name for this kind of recommendation 15:50:52 yeah, I don't think anything uses that word in the game menu 15:51:00 it does 15:51:15 ah I see it 15:51:25 that's probably because the word is short 15:51:33 yeah, seems plausible 15:51:38 "Recommended" might be too long? 15:51:40 not sure how that menu works 15:51:51 <|amethys1> well, and the old system used to be more about "viable" 15:52:02 re: v.1, keep in mind that this species is actually similar to the current gnoll 15:52:09 -!- |amethys1 is now known as |amethyst 15:52:30 I believe the warrior mage class recommendations about the fact that gnolls have good spell skill and good melee skills necessarilly 15:53:05 so I think they definitely sense, but that doesn't mean we have to recommend all of them 15:53:18 since we want to have some variety in the recommendations 15:53:22 *definitely make sense 15:53:56 we recommend 4 out of the 5 warrior mages which does seem high 15:55:11 one challenge is that once you start picking the regular book starts it's not obvious how to choose among them, so maybe adding wz to the list covers that ground 15:55:47 right, if another was chosen, probably the best choices would be Ne Su IE 15:56:21 you just want to think about which books starts are more likely to incorporate a good deal of melee due to having spells that more strongly synergize with this 15:56:34 but those three I named probably get recommended a lot already 15:56:37 I don't think gn recommending wz requires wz recommending gn, right? that's what |amethyst was alluding to 15:56:43 <|amethyst> right 15:56:44 yeah, that's right 15:56:55 <|amethyst> hm 15:57:11 <|amethyst> the newish races don't seem to be recommended by any jobs 15:57:26 <|amethyst> so it would probably be worth tweaking that too, where appropriate 15:57:46 yes 15:58:07 Ne only recommends 5 15:58:23 woah 15:58:31 I just crashed, now how did I do that 15:58:42 -!- Mindiell_ is now known as Mindiell 16:01:04 http://dpaste.com/2AHT5JN 16:01:26 seems to have started a summoner and caused a crash 16:01:26 let me see if I can recreate 16:01:31 -!- amalloy_ is now known as amalloy 16:02:31 started a ddsu sort of by accident and it crashed 16:02:43 huh, I've seen that crash a few times but had no idea it was game beginning like that 16:03:06 ah 16:03:06 advil: can you try choosing summoner first then DD? 16:03:06 I just crashed again 16:03:10 one sec 16:03:53 hrm, it's not consistent I guess 16:04:38 yeah I can't seem to cause it again, but I did get a second crash once 16:05:51 fwiw there are a lot of races recommending Wz already. 9 of them, compared to 4 or 5 for stuff like AM/Sk 16:05:51 looks like it's the same type of crash the second time 16:06:18 not to say that it's wrong to have yet another race recommending Wz 16:06:45 haven't gotten it so far...my laptop is a bit behind master 16:06:45 I can't seem to reliably recreate anyhow 16:07:19 the best kind of bug, one that only triggers 2% of the time 16:10:47 ok, crash again, let me see if I can do this 16:11:02 gosh darn it 16:11:19 the stack trace makes it look like something that could happen any time you pick up a book, which is guaranteed to happen on turn 0 by a book start 16:12:59 advil: i think ShoppingList::ShoppingList() just doesn't initialize its vector member, and assumes that means it will be nullptr but actually it could be anything 16:13:22 haha nice 16:13:26 why does it even have a CrawlVector* instead of a CrawlVector? 16:13:37 what even is a CrawlVector 16:14:14 a vector that tags.cc can automatically serialize, and iirc it has tagged unions for values, rather than being homogenous 16:14:50 I'm suprised that advil hasn't run into the infamous CrawlVector before 16:15:04 me too 16:15:06 amalloy: that's a surprisingly reasonable answer 16:15:07 also glad that amalloy gave that summary because mine would have been terrible 16:15:11 I think I have, but I forgot about it 16:15:28 it's CrawlHashTable, but indexed by integers 16:16:17 it wouldn't be CrawlCode without a CrawlVector 16:16:19 let me try to repro the crash locally, and then i'll try a simple fix 16:16:31 we can put a real fix in later. i just have to start a DDSu? 16:16:50 I seemed to get it by choosing Summoner then Deep Dwarf 16:17:19 but based on what's been said I doubt order is important 16:17:21 and it only happens like one time in ten from what I've seen 16:17:39 I still can't replicate it, but amalloy's hypothesis would probably lead one to expect different behavior on different compilers 16:17:41 cool, repro on the third try 16:19:17 second try after applying my "fix". lovely 16:19:23 heh 16:19:34 oh I just got it 16:19:46 hm, maybe this is different 16:19:59 I triggered a different ASSERT at least 16:20:23 and this was a segfault 16:20:39 oh yeah, i'm getting a different assert than gammafunk was too 16:20:47 i'll try it in the debugger 16:20:55 blah blah SFLAG_UNSET? 16:21:26 yeah 16:21:29 it's roughly in the same place 16:22:38 oh, and now i get the one gammafunk was getting, while in the debugger 16:26:19 aidanh: you may be interested, as this seems to be code from: 16:26:19 %git 4f8b0bdfded 16:26:19 07Aidan Holm02 * 0.22-a0-382-g4f8b0bd: Clear useless books from $-list after finding spells 10(3 weeks ago, 3 files, 37+ 3-) 13https://github.com/crawl/crawl/commit/4f8b0bdfdedd 16:26:39 I was having this issue on the ui-overhaul branch but that was ~ a week ago. It was happening every time I loaded a char with a level 1 spellbook. 16:29:27 ok who wrote spells_added_to_library :-P 16:29:49 also who has the @crawlcode password these days 16:30:27 if (!thing_is_item(thing)) // ??? 16:30:27 continue; 16:30:34 is my candidate 16:30:46 tip top stuff :D 16:31:40 though probably this is modeled after something else, maybe cull_identical_items -- at least whoever wrote this added the ??? 16:32:16 advil: aidanh added this function in the commit i linked above 16:34:42 amalloy: was your first hypothesis about `list`? 16:34:51 yes 16:35:10 what if we had shops where you could buy advice 16:35:20 i think ShoppingList::refresh() is being called when the crash doesn't occur, and is not being called when the crash does occur 16:36:26 further, anecdotally, it seems that the crash occurs when i start a second game of crawl without exiting all the way to the shell. ie, start a game, ^Qyes, then from the menu start another game 16:36:58 there is some leftover stale state from the first game that causes the setup for the second game to break, maybe in you.props 16:37:07 ah, sounds plausible 16:37:29 though if refresh fixes it, probably not in you.props 16:37:41 refresh reads from you.props 16:38:31 yes, but only resets the list if that key doesn't exist 16:38:31 fwiw I was not reliably getting it when I did ctrl-q to menu, though 16:38:50 yeah, i think i see what you mean. i'll try to figure out what codepath leads to calling refresh and see why it might not go there the second time 16:50:10 well, this seems to fix it but i'm not thrilled, nor am i sure it won't break some other weird case. i'll push my fix but if anyone can find a better place to put this call to refresh i'm happy 16:53:25 -!- pauldub_ is now known as pauldub 16:53:26 -!- neizenel_ is now known as neizenel 16:53:28 -!- bytesighs_ is now known as bytesighs 16:53:47 -!- yaknyasn_ is now known as yaknyasn 16:53:49 -!- mroth_ is now known as mroth 16:54:33 03amalloy02 07* 0.22-a0-404-g35e5176: Fix a crash in game startup 10(3 minutes ago, 1 file, 4+ 0-) 13https://github.com/crawl/crawl/commit/35e51767bf18 16:58:14 i see now why ShoppingList::list is a pointer now though. it doesn't actually want to own the vector; it wants to be an alias for you.props[SHOPPING_LIST_KEY] so that edits through ShoppingList are automatically propagaged to you.props, from which they are automatically (de)serialised 16:59:58 i guess that would be more useful as a comment than an IRC message 17:01:26 -!- Rjs- is now known as Rjs 17:01:26 yeah I guess that's a very common pattern 17:01:51 using an alias to the prop so that changes remain permanent 17:05:11 ^status 17:05:11 43 Crawlers. CBRO disk usage=99% (135GB) | RAM usage=51% (4GB)| uptime/CPU= 17:05:01 up 120 days, 20:40, 3 users, load average: 0.68, 0.78, 0.88 (4 Cores) http://status.berotato.org 17:05:14 johnstein: looks like we're almost full on cbro again 17:05:38 ah, and i understand the cause a little better new. in game setup, we write `you = player();`, which clears the props table. but we haven't cleared the shopping list, and it still refers to the memory that used to hold your shopping list, which might or might not be valid 17:06:52 hm 17:07:03 aidanh might have found this already, he was running some stuff through valgrind 17:08:05 and I remember he found something 17:08:13 this is sorta a questionable use for a prop imo, since there is only one function that ever writes or reads that prop. it's "convenient" to not have to write the tags.cc code, except that actually it wound up causing a bug. i think it would be cleaner for the vector to live inside the shopping list, and persist it by hand in tags.cc 17:09:31 yeah 17:09:50 Unstable branch on crawl.jorgrun.rocks updated to: 0.22-a0-404-g35e51767bf (34) 17:10:26 that's not what he found in valgrind btw, or at least not what he mentioned finding 17:12:10 the shopping list seems like it should be like a vector in player 17:12:23 well, there's a lot of stuff in player 17:12:30 i understand not adding more and more fields 17:12:50 but it's now implciitly a field on player 17:12:55 because it's in you.props 17:12:59 *implicitly 17:13:07 but yes, it would be nice if the canonical place were inside of player, or some other useful class, instead of being a top-level singleton 17:14:22 you.hopes_and_dreams.shopping_list 17:18:06 looks like the shopping list as we know it today was added in 2009, and its storage has been the same ever since: a global singleton object, with a pointer member aliasing you.props 17:22:56 advil: i'm in favour of moving it to a field of player, but i'm not excited about doing it myself. are you interested? 17:23:06 haha 17:23:13 I can add it to the list 17:23:59 it would involve a bunch of tedious worrying about save compat, depending on how we do it 17:24:10 though I suppose it could just set the prop before serializing 17:53:39 what does the dev workflow look like these days; it seems the wiki is no longer used? 18:00:35 we've been using the github wiki more 18:01:00 the doku wiki is the public one that you can still sketch proposals on 18:01:25 or you can make a wiki page on your own fork of the the repo and give us a link (which we can add to planning pages) 18:01:43 but most changes just happen in local branches and are not planned out via wiki, same as before 18:09:44 splendid 18:09:57 I'm not sure if I have ideas but I have The Itch again 18:10:34 I'd see a doctor about that if I were you 18:15:37 nah I'll just e-mail patches 18:19:36 the secret dream would be an octopus king wizlab but I doubt octopode crushers would be brought back for that 18:20:35 Unstable branch on underhound.eu updated to: 0.22-a0-404-g35e51767bf (34) 18:21:54 -!- Krakhan|2 is now known as Krakhan 18:37:41 ebering: you must defeat me, the creator of octopode crushers, before you can summon them from the grave 18:37:48 they answer only to their living master... 18:38:13 iron giants with an octopus tile? 18:38:26 they predate iron giants!!!! 18:38:34 well, in main crawl they do 18:38:53 iron giants were/are a sprint monster before grunt created them 18:39:11 I'm pretty happy with the harpoon shot thing since it solves most of the problem that the throw thing did 18:39:30 and it's a vastly simpler implementation to boot 18:39:43 although I think much of that code, much refactored by PleasingFungus and wheals, lives on with robin's throw ability 18:39:52 oh and iron giants can still throw monsters 18:40:08 .harpoon 18:40:10 57. Kairos the Cleaver (L11 VSFi of Okawaru), slain by a swamp worm (kmap: worms_lemuel) on Lair:1 on 2018-05-07 15:43:49, with 7682 points after 6538 turns and 0:18:08. 18:40:14 .harpoon cikiller=iron_giant 18:40:16 2. TheAlmightyOne the Bringer of Light (L27 HOMo of The Shining One), blasted by an iron giant (iron shot) on Dis:7 (iron_city_of_dis; dis_wall_7) on 2018-04-07 23:44:30, with 865522 points after 70685 turns and 6:16:59. 18:40:32 so far only two harpoon-based iron giant kills, and I don't think that was a major factor in either death 18:40:39 .harpoon s=ckaux 18:40:40 57 games for * (harpoon cikiller=swamp_worm|iron_giant): 36x, 19x harpoon shot, iron shot, some deep water 18:40:45 nice 18:40:50 a collision death 18:41:00 .harpoon ckaux=some_deep_water 18:41:01 1. Steelmeadow the Spry (L17 HaAs of Dithmenos), collided with some deep water caused by a swamp worm on Swamp:4 on 2018-05-05 02:28:51, with 131006 points after 45952 turns and 2:47:47. 18:41:07 .harpoon ckaux=some_deep_water -tv 18:41:08 1. Steelmeadow, XL17 HaAs, T:45952 requested for FooTV: telnet://termcast.develz.org or http://termcast.develz.org. 18:41:19 harpoon shot spooked me playing earlier today 18:41:36 haha, wow 18:41:38 good tele 18:41:44 ya 18:41:49 oh wow 18:41:52 another good tele 18:42:17 yeah it's pretty spooky, I'm sure it'll cause a lot of collateral kills that won't be easy to find with queries 18:42:36 I've cleared swamp a few times with them and I had one dangerous incident that cost me a blink scroll 18:42:46 but I didn't actually come close to dying 18:42:50 so the player doesn't get thrown anymore though :( 18:42:56 yes 18:43:36 but the reason why I like pull better than push is that the push introduces this silly thing where if you're trying to melee the monster, you have to sort of keep trying 18:43:51 which it wasn't very safe to melee monsters that had this, but it was particularly annoying 18:44:10 and the throw effect had a lot of randomness inherent to it that made any kind of planning a reaction difficult 18:44:49 the pull mechanic is more predictable and still a reasonably strong effect that works normally in terms of dodging and blocking 18:45:17 also ebering, players don't get thrown but they can collide with things! 18:45:25 mostly deep water or lava if they're not flying 18:45:45 but possibly monsters if a shot misses one of your allies and hits you, obviously that's a corner case 18:46:14 and I like that there's this tradeoff with flying to prevent said collisions 18:46:23 if you fly, you'll get pulled further but take no collision damage 18:46:52 assume e.g. the swamp worm is in deep water 18:47:03 wouldn't be too relevant for iron giants 18:47:16 it would if they were octopodes 18:47:26 ! 18:48:23 def. harpoon shot is very cool and produces an effect similar to but more fun than player throw 18:49:59 ebering: you could make an octopode unique 18:50:48 I don't think we want to make too many "one-off" species monsters since those end up being hard to make fit into the regular game in a meaningful way and just start from a flavor basis 18:50:52 but uniques can be more exceptional 18:51:16 people have spitballed some kind of "ring master" unique but I'm not sure I've ever heard a good proposal for an op unique yet 18:53:13 yeah I'm not sure I have any good ideas 18:54:39 the joke one is an op w/malign gateway and a throw attack that sends the player to next to the tentacle but that seems too mean to fight 18:57:02 and also my memory is that MG isn't that scary as a monster spell, I've seen it on some liches and you can just walk away 18:58:07 ebering: the op throws the MG portal near the player, like it's a summon bomb 18:58:16 (that's a silly idea) 18:58:56 I'm not too keen on reviving throw attack in general since it was really complicated for not great gameplay 18:59:08 it even has to have its own ktype because it does damage so differently 18:59:17 s/has/had/ 19:17:57 TIL about harpoon shot. cool spell 19:19:35 it allowed us to finally stop swamp worms from submerging 19:19:41 but I think we have to remove that submerging code 19:19:54 you'll never find it. it's submerged 19:19:54 kraken tentacles are the only thing doing this currently, and they don't need it 19:20:01 .....i don't get..... 19:36:35 Hello to anyone who is on board! I have another question about species editin. I want to make it so that my dude gains +1 to two stats every other level, rather than one. Can I simply tweak around the syntax in species-data or do I need to code such thing from scratc, like with demigods ad their double chosen stat? 19:38:46 It doesn't look like there's anything you can do in species-data.h to that effect 19:41:33 Aw jeez, that's a pity - will leech from demigod statgain then 19:43:51 -!- Pinkbeas1 is now known as Pinkbeast 19:44:02 Another one - is there a cap on experience aptitude? Jugding from Gnolls - there is no for skill aptitudes, but what if I set the experience to more than 1? Will it even matter? Didn't notice much difference when cranked it up to 3. 19:44:44 you might get weird behavior at extremely high/low values, depending on how that stuff is actually set up (too lazy to check) 19:45:35 the current range among crawl species is only from +1 to -2, I'd imagine that you can get quite a bit outside that before running into any problems 19:49:10 I see, thank you 19:49:42 Are you that hellmonk who does hellcrawl fork? 19:49:54 yes 19:51:21 Nice! Trying to get around dcss tinkering precisely because I had a great time with your fork and now want to do some modding myself. Thanks 19:59:41 Sergeek: you can confirm that your experience aptitude is working by playing around in wizmode. start a HuFi, create a golden dragon, and kill it. see what level you get to, and then try again with your new XXFi 20:00:19 03advil02 07* 0.22-a0-405-g6463fda: Handle startup preferences (esp. name) better when restarting 10(3 minutes ago, 1 file, 3+ 2-) 13https://github.com/crawl/crawl/commit/6463fda370fe 20:00:36 I still haven't replicated those crashes on my home computer (they must just be rarer on mac for some reason) but I did find some other bugs 20:02:38 oh haha amalloy already fixed it 20:03:43 like hellmonk, i would expect it to basically work 20:09:52 Unstable branch on crawl.jorgrun.rocks updated to: 0.22-a0-405-g6463fda370 (34) 20:32:40 urgh, yeah, that shopping list code is a mess 20:35:04 I guess I'll have to add valgrinding multiple games in a row to my pre-push checklist 20:36:17 the valgrind bug I found before was a write() of uninitialized memory while saving chunks, but I was never able to figure that one out 20:37:50 Currency, if you have any feedback on the ui-overhaul branch changes, I'm all ears 20:47:44 aidanh: I just ran into that bug before. Didn't mean to give offense. I was referring to the silly boolean expression naming rather than the bug itself but I get that's probably not clear. 20:52:15 No offense taken at all! I ask because you're the only person who's mentioned testing that branch out :) 20:54:22 Yeah, thing_is_item() is pretty bad; I threw it in for save compat reasons, but it's be better to redo shopping list storage entirely as others discussed 20:55:09 Oh. My apologies then. Sure. I've been playing with it a lot. My personal pet peeve I've been trying to crack is getting CMD-Q to function with it on OSX. 21:00:17 I tried to send you a private message about bug reports for the branch/help with it at one point but I suppose you never got it. I figured manits was cluttered enough without feature branch issues. 21:05:38 I'm afraid not; if you tried to send it over Irc it's possible my bouncer lost it 21:06:45 if you mention me here in crawl-dev I'll see it though 21:07:05 Ah ok. I did ./msg 21:07:52 any chance your client logged it? eager for bug reports here :) 21:08:33 also, is the cmd-q thing specific to the ui-overhaul branch? 21:10:08 %git 7aac3cfdd960 21:10:08 07wheals02 * 0.18-a0-1763-g7aac3cf: Fix autofight if you remap keys (#5634). 10(2 years, 1 month ago, 2 files, 102+ 26-) 13https://github.com/crawl/crawl/commit/7aac3cfdd960 21:10:15 Yeah. The normal game handles it and saves the game. 21:11:47 It's not clear to me how yet. I'm really unfamiliar with how C++ does os level stuff like this beyond basic atexit() 21:14:15 hrm, both autofight and automagic use a delta_to_cmd (formerly delta_to_vi) that they define themselves 21:14:31 seems like something that could really be an appropriate clua lib 21:14:49 er, not formerly, since I have a patch to rename it 21:19:02 Currency, thanks, I'll look into that; crawl uses SDL2 for windowing stuff like that, so it's probably an error with quit event handling 21:30:35 can you modify e.g. the crawl lua metatable from a lua file? 21:31:00 I'd like to put delta_to_vi (soon to be delta_to_cmd) in that metatable but I don't really want to rewrite it in C++ 21:31:09 since it seems fine to have as lua 21:32:36 I guess there's the question still of where I'd modify it from in terms of a lua file 21:32:54 iirc you should be able to get the metatable and modify it, but not assign a new one to crawl 21:33:29 unless your metatable has a metametatable preventing modification, of course 21:33:53 yeah, sounds reasonable, although I realise that this makes things a bit weird in terms of where I'd put this modification 21:34:15 right now we load a bunch of files in dat/clua at game start at some point 21:34:46 they are loaded in order, so I could this in e.g. the first file in the list, but this introduces an order dependency 21:34:46 which isn't presently there 21:35:00 !source initfile.cc:1571 21:35:00 https://github.com/crawl/crawl/blob/master/crawl-ref/source/initfile.cc#L1571 21:35:21 alternately I could simply put this in a lua file and not add this to the list 21:35:22 -!- amalloy is now known as amalloy_ 21:35:31 and then have the relevant lua files load it 21:36:55 perhaps this is how dlua stuff is already basically handled 21:37:13 only explicitly load a few files and have other files in the same dir that are sourced by said files 21:37:49 alternatively, yeah, it might work, but would you be proud of the work? 21:39:06 I'll be proud when I send a DMCA notice to take down pizza tornado 21:42:13 the way luakit does it is to have a core c library and then additional helper libraries written in lua that can be require()-ed in 21:42:33 <|amethyst> gammafunk: when you say "modify the metatable", do you mean that you want this to be crawl.delta_to_vi instead of just delta_to_vi ? 21:42:42 |amethyst: yeah that's what I was thinking 21:42:46 maybe that's just not necessary 21:42:50 but then crawl seems to have disabled a lot of builtin lua stuff like require, so I'm not sure if that'd work 21:43:21 aidanh: yeah looking at dlua.cc, dlua just loads dungeon.lua specifically and that starts a chain of includes for all the other files 21:43:25 <|amethyst> gammafunk: that part's easy enough ignoring the dependency issues: crawl.delta_to_vi = function(dx, dy) ... 21:43:27 crawl.delta_to_vi would be better for namespacing purposes at least 21:43:41 ok 21:43:50 so I'll probably just make a lua file in clua 21:43:59 and not add it to the initfile.cc load list 21:44:07 and have autofight and automagic er..no 21:44:10 I don't need to do that 21:44:17 I'll just define it into crawl.delta_to_cmd 21:44:30 and then modify those files to call that 21:44:42 at least I hope the crawl lib is defined by the time those are loaded 21:45:15 <|amethyst> it doesn't have to be defined before the users are defined 21:45:15 <|amethyst> just before they are run 21:45:17 gotcha 21:45:25 I didn't think about how lua does scoping 21:45:55 lexical? I've already forgotten that bit of CS 21:46:25 <|amethyst> it's still lexical scoping, but it doesn't check for undefined variables at compile time 21:46:35 <|amethyst> python e.g. is similar 21:55:44 hrm, I guess I do want to add this to the initfile list 21:55:49 since it would be weird to include it twice 21:56:11 as it's modifying the crawl table 21:57:00 <|amethyst> could put it into userbase.lua 21:58:10 <|amethyst> (it's in the dlua directory, but is loaded automatically into the clua interpreter, see CLua::init_lua 21:58:13 <|amethyst> ) 21:58:28 that sounds good, yeah 21:58:45 -!- amalloy_ is now known as amalloy 22:01:18 well, it looks like my plan is temporarily thwarted by the fact that I actually have to rewrite part of automagic 22:01:45 since this new delta_to_vi function works differently 22:02:00 so I'll just push this rename patch for autofight for now and look to clean that up later 22:08:47 it's like 7 hours late for the discussion on it, but all the Gn recommendations were placeholders and I never got around to fixing them 22:09:04 I brought it up a couple times, but nobody seemed to want to discuss it so I just forgot about it 22:12:24 just rec them every mage bg tbh 22:16:41 yeah I don't think recommending warrior-mages is bad, those classes make sense for Gn, but probably we don't need to recommend quite so many 22:17:34 Ar would make sense to me, obviously Wz, could add in something like Ne or IE 22:18:35 I think they were originally recommended Wz, but some people complained (Gn v1.0) so I swapped it to Wn in spite 22:18:37 and it could lose one or two of the warrior mages 22:18:42 haha 22:18:57 Wn is still good, so we wouldn't change that one 22:19:49 elliptic: by the way, if you have any opinions on what classes Gn should recommend or which classes Gn recommend, we'd value your input since you did still the current idea from mikee 22:19:53 also, I think I forgot to any recommendations the other way around: picking a class first will never recommend Gn 22:19:55 s/still/steal/ 22:20:02 right, also a problem for Ba? 22:20:24 probably 22:39:51 it seems like the webpage "live games" listing is only cbro and cpo these days? 22:47:07 !bug 11403 22:47:07 https://crawl.develz.org/mantis/view.php?id=11403 22:47:25 gammafunk: while you're at it, you might as well merge the patch in 11403 22:47:46 I would have done that already but I forgot about it because it's in mantis 22:56:54 advil: that's exactly what I'm merging, yeah 22:57:32 I am just making a commiter's note about the automagic issue and I will ping w h e a l s about it after I push in case he wants to fix automagic himself 23:02:55 ah ok 23:03:58 03Poroing02 {gammafunk} 07* 0.22-a0-406-g9ff8d77: Fix obsolete function name. 10(2 months ago, 1 file, 7+ 7-) 13https://github.com/crawl/crawl/commit/9ff8d774c118 23:05:03 wheals: I made a note about in the message of 9ff8d77 about updating automagic to have the same kind of change to the (former) delta_to_vi() function you made for autofight a long time ago. 23:05:36 wheals: if you wanted to update automagic since you might understand the relevant code still, please feel free, otherwise I can take a look at that later 23:07:22 oh woah, splash screen request 23:07:33 ah yeah, this is from dynast 23:08:02 https://crawl.develz.org/mantis/file_download.php?file_id=7634&type=bug 23:09:26 oh right, there are still some code issues for using all of roctavians tile variants 23:10:29 advil: is the bug mentioned in https://crawl.develz.org/mantis/view.php?id=11268 fully fixed by now? 23:10:49 I guess there's still an open PR on our sdl contrib, maybe two 23:11:03 oh no there's just one PR 23:11:18 oh I never even saw tha tPR 23:11:42 I merged something very similar 23:12:56 free bug report fixing credit if you want to resolve that one 23:16:32 I'm not 100% sure if his change covers the same cases 23:59:36 Unstable branch on CRAWL.XTAHUA.COM updated to: 0.22-a0-406-g9ff8d77 (34)