01:11:21 PF: not trying to police it, just make suggestions, kind of thought sefie would reply going eg "what's new Tavern" 03:27:00 <[808]state> Are there any tasks/issues that the devs would like help with? 03:33:38 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-5208-geafff8c3b6 05:06:42 Unstable branch on crawl.akrasiac.org updated to: 0.33-a0-224-g21dfed3 (34) 07:55:24 <08nicolae> q: when levelgen fixes up a vault without no_pool_fixup, and replaces w with W, does that occur during any point that can be affected by the .des file. 07:55:48 <08nicolae> i feel like the answer is probably "no, pool fixup happens after placement and it's too late for the des to do anything" but if the answer is somehow yes that will be good to know 07:59:10 <05icoson> definitely gotta keep an eye on the .io thing, hope that all the rich companies that use io domains pressure to keep it 07:59:27 <05icoson> sounds like a minimum 5 year transition period at least 08:43:39 <06mumra> That's correct, it happens in _prepare_water which happens after all vault placement and even after _post_vault_build 08:46:09 <06mumra> I was going to suggest the easiest way to do something like you describe is to expose _prepare_water to dlua -- so you can call that yourself from lua script. But it operates on the actual map grid, I'm not sure that works even in a des file, since the map grid isn't even written yet 08:47:27 <06mumra> But the logic of _prepare_water is actually not all that complicated and could be replicated in Lua fairly trivially, depending what you want to do with it 08:54:48 <06mumra> (Looking at _prepare_water I think the algorithm is a bit buggy, well it has a directional bias depending on the order the iterators run in ... certain squares are either more or less favoured to be fixed up along N/S/E/W edges I think...) 10:48:33 <08nicolae> interesting... 🤔 10:58:22 <06mumra> In fact, reading it again, i think the effect is more pronounced than that. Because it is iterating every grid, and then checking what the adjacent squares are currently, and modifying it as it goes. So each change will affect subsequent iterations. So depending on the orientation of your water feature you could end up with a larger section of shallow water generated. I'm just not 100% which order the iterators will go over tiles in. 10:58:22 Suppose you have www... ...www www... vs ...www www... ...www For the left-hand one it's only possibly to get at most wwW... wwW... wwW... Whereas the right-hand one it could spread much further and get like ...WWw ...WWW ...WWw 10:59:02 <06mumra> Assuming the rectangle iterator goes left-to-right (which I think it does). And there will be a top-to-bottom effect as well. 10:59:34 <06mumra> So maybe the left-hand one has a chance of spreading down and left wwW... wWW... WWW... 11:00:29 <06mumra> Basically: it's a bit weird and probably not working as intended. Would be better to build a vector of coord_defs that are going to get updated, and then actually update the grids in a separate loop. 11:39:04 03DracoOmega02 07* 0.33-a0-225-g19655175bd: Fix Wucad Mu / Borg wizlab summons being broken (midn8) 10(12 minutes ago, 1 file, 9+ 1-) 13https://github.com/crawl/crawl/commit/19655175bd54 11:47:39 <08nicolae> about how hard would this be to implement, do you think. asking for a friend who is thinking about vault-related code projects to work on in the near future 11:48:34 <04dracoomega> I think an important question is actually: what do you want to do? 11:49:39 <04dracoomega> Probably easier to say how difficult something is (or the what would be the best way of going about it) if it was more clear what you wanted to do with inserting some vault code in the middle there 12:13:43 <08nicolae> the original question was along the lines of "if i wanted to have a chance for shallow water to have a monster or something in it, would i need to do set up the shallow water manually, or would it be possible to let pool fixup happen and then piggyback off that", i assumed it was the former, and it was. i don't actually need to mess with the prepare_water algorithm itself for any vault purposes, it's just that knowing about weird 12:13:43 levelgen inconsistencies (the one mumra pointed out about how the iterators work) drives me very slightly crazy, as if i had read a very tiny and inconsequential necronomicon. (see also: big-V Vaults vaults placement not rotating things correctly) i hope they get fixed but they're pretty low-key issues that really only stand out if you have vaultsbrains so if i want it done i'd either have to do it myself or wait for someone who knows the code to finish 12:13:43 all their own stuff 12:20:59 <04dracoomega> Makes some sense to me 12:21:58 <04dracoomega> (I guess if you tried to place a non-swimming/flying monster on deep water via a vault, it would either not place it or replace the floor beneath with dry land entirely, huh?) 12:22:26 <08nicolae> i forget what happens in that case, i'll have to try when i get home 12:25:28 <08nicolae> in any case, the original question is by no means insurmountable, it just takes a lil more work. such is life, as a vault designer. 12:26:03 <08nicolae> in related news though i've got a lot of lair decors for r-i 12:27:11 <04dracoomega> Nice 12:28:14 <08nicolae> more than once this week i was thinking "okay, i've got a good number, when i get home i will make a PR" and then ten minutes later i think of another 12:50:39 03DracoOmega02 07* 0.33-a0-226-g4eba2d974d: Fix mangrove hornets being marked as 'minions' (Riverefiend) 10(83 seconds ago, 1 file, 9+ 10-) 13https://github.com/crawl/crawl/commit/4eba2d974d75 14:18:57 <12ge0ff> When playing on CXC, if you leave the game open for more than 5 minutes w/o pressing any keys, you'll get disconnected. This doesn't happen to me on any other server, but it does happen on CXC to other players too. Is there some server setting for dealing with stuck games? If so, would it be possible to make it less aggressive? 14:19:08 <12ge0ff> ^ @melvinkitnick 15:38:53 <09gammafunk> &versions 15:39:16 <09gammafunk> ??rebuild 15:39:17 <04Cerebot> rebuild[1/2]: Devs can trigger rebuilds for these servers: https://crawl.akrasiac.org/rebuild https://cbro.berotato.org/rebuild https://crawl.dcss.io/rebuild https://archive.nemelex.cards/rebuild Bug gammafunk, advil, or Nap.Kin for CDO. Use your powers wisely. 15:39:20 <04Cerebot> CAO: 0.33-a0-224-g21dfed3, CBR2: 0.33-a0-224-g21dfed3d1e, CDI: 0.33-a0-226-g4eba2d974d, CDO: 0.33-a0-224-g21dfed3d1e, CNC: 0.33-a0-226-g4eba2d974d, CPO: 0.33-a0-218-g4438645, CUE: 0.33-a0-224-g21dfed3d1e, CXC: 0.33-a0-163-g6f3bd40977, LLD: 0.32.1 15:39:42 <09gammafunk> @melvinkitnick probably not related to Ge0ff's issue, but Flugkiller is reporting that cxc hasn't updated in a week, looks like it's still at 0.33-a0-163-g6f3bd40977 15:40:01 Unstable branch on underhound.eu updated to: 0.33-a0-226-g4eba2d974d (34) 15:49:06 <09gammafunk> For https://github.com/crawl/crawl/issues/4067 I'm included to treat this as a feature (of felids) rather than a bug 15:49:24 <09gammafunk> This person is just reporting that a revived felid can loot multiple paths in a Gauntlet 15:50:00 <04melvinkitnick> I’ll look both issues tomorrow and let you know. 15:50:07 <09gammafunk> thanks a lot! 15:51:02 <09gammafunk> Felid lives are valuable consumables. It seems pretty much in the design of felid to allow this type of gameplay, but I guess I can see some argument that purposefully killing yourself for more gauntlet loot is not great gameplay, if you happen to be confident that you'll have enough lives in the near term before you gain another 15:51:29 <09gammafunk> I was going to comment along the lines of "this isn't a bug" and then close the issue, but if any other dev feels strongly, let me know 15:51:44 <09gammafunk> (or respond in the issue) 15:57:32 <12ge0ff> felids respawning in the gauntlet is definitely fine and doesn't require any fixes 15:58:56 <12ge0ff> > purposefully killing yourself for more gauntlet loot is not great gameplay it can look like "i'll try fighting the minotaur, but i won't spend a bunch consumables in the fight, so even if i'll die, i'll be able to loot another path", which also feels fine 16:14:23 <04dracoomega> Yeah, I felt similarly when I read it. 16:14:41 <08nicolae> felids should know only misery and privation. a felid who dies in the gauntlet appears in a special zone off to the side called the "reanimation chamber" where they are confronted by eleven minotaurs and a twelfth minotaur with fourteen arms 16:15:54 <04dracoomega> (I'm not convinced that deliberately trading a life for more stuff is often even correct, but 'choosing to sacrifice a life in the pursuit of some other goal' just feels like normal felid play - for instance, deciding whether to risk wasting consumables on an unlikely escape from death as opposed to saving them for a more reliable opportunity later. If anything, this is a cute interaction.) 16:16:29 <04dracoomega> Like, a gauntlet is a place where you're more likely than average to need that extra life if you're wanting to do a bunch of additional fights in it, imo 16:17:13 <12ge0ff> (also, if in the future xp evokers become less unique, there will be less incentive to intentionally splat in the Gauntlet) 16:43:10 <06pleasingfungus> yeah, seems fine and good 16:43:15 <06pleasingfungus> more dead cats please 18:45:58 <06mumra> What's actually the problem with V rotation? I remember spending some time trying to get that right when I did all that ... 19:22:58 <06mumra> https://cdn.discordapp.com/attachments/747522859361894521/1294123046885392489/image.png?ex=6709dd82&is=67088c02&hm=bd4ce911e73b065d8bdf059921fe7cf9952e6fb747ff66e1459609d9116f3d19& 19:23:12 <06mumra> Did some testing and confirmed the suspected behaviour with pool fixup 19:23:39 <06mumra> (Vault is dpeg_arrival_castle and it's all deep water in the des) 19:24:05 <06mumra> So the shallow water is "spreading" SE from the top and left edges; at the bottom right it's just 1 tile thick 19:31:26 <06mumra> After fixing the algorithm to be more fair, it's just a one tile border of shallow water all around 19:31:28 <06mumra> https://cdn.discordapp.com/attachments/747522859361894521/1294125185145901086/image.png?ex=6709df7f&is=67088dff&hm=4caba6afc096e0469bff48ae3bdab6a2261d69dea2c3572833807bc0fe26c106& 19:32:45 <04dracoomega> The directional bias is an unfortunate artefact, but I can't help but wonder if 'never more than 1 tile in' is actually the ideal look here? 19:33:05 <06mumra> Yes i prefer the look where the shallow water "spreads" a bit further, so I tried just running 2 passes of _prepare_water and it makes a result more like this 19:33:07 <04dracoomega> (Not that this is something I've given much conscious thought to before) 19:33:07 <06mumra> https://cdn.discordapp.com/attachments/747522859361894521/1294125599811698710/image.png?ex=6709dfe2&is=67088e62&hm=e897d07b8fc984cd65da67c4033af5bc9f5fbc474717c8e67d04da88b3be3de9& 19:33:50 <04dracoomega> Hmmm... I wonder if that's too likely to consistently make the total perimeter shallow? 19:34:24 <04dracoomega> (Maybe that's fine. I'm just wondering out loud) 19:37:40 <06mumra> Yeah you're right tho... can tweak the probabilities a bit 19:39:55 -!- The topic of #crawl-dev is: Crawl Development | https://github.com/crawl/crawl | Logs: http://s-z.org/crawl-dev/, temporarily http://crawl.akrasiac.org/logs/cheibriados/ | People with +v have commit access, devs on bridged discord as well | General Crawl-related chat to #crawl | Long stuff to a pastebin service, please 19:39:56 -!- The topic of #crawl is: Play Dungeon Crawl Stone Soup online now! Type ??online for instructions, ??lg / !lg for play stats | PM Sequell for long queries | http://crawl.develz.org | FooTV game replays: ??footv for instructions | #crawl-dev for dev discussion, #crawl-offtopic for offtopic 19:44:47 <06mumra> https://cdn.discordapp.com/attachments/747522859361894521/1294128538357600256/image.png?ex=6709e29f&is=6708911f&hm=4240adc9b504342965c16f965a31dbb4169b9bd1bb15c5b9a5bd16e1116a642e& 19:47:47 <04dracoomega> Nicer 20:55:53 03mumra02 07* 0.33-a0-227-g7f099897d9: Fix directional bias in pool fixup 10(13 minutes ago, 1 file, 9+ 3-) 13https://github.com/crawl/crawl/commit/7f099897d982 21:08:50 I found this part in melee-attack.cc. Does someone know what it does, especially why apply_damage_brand() is called twice for chaotic weapons? 21:08:51 apply_damage_brand(); 21:08:51 if (weapon && testbits(weapon->flags, ISFLAG_CHAOTIC) 21:08:52 && defender->alive()) 21:08:52 { 21:08:53 unwind_var save_brand(damage_brand); 21:08:53 damage_brand = SPWPN_CHAOS; 21:08:54 apply_damage_brand(); 21:08:54 } 21:11:33 <06pleasingfungus> !source melee-attack.cc 21:11:34 <04Cerebot> https://github.com/crawl/crawl/blob/master/crawl-ref/source/melee-attack.cc 21:13:30 <06pleasingfungus> sefie: https://github.com/crawl/crawl/commit/c468895e6595af899960b6686c9c36dd479a2e56 21:24:47 mhm, thats good advice, to look through the history and check for changes of the code in question. Thanks! 21:28:47 <08nicolae> i'd have to remember the specifics but i had some vaults that wouldn't rotate. or maybe it was they couldn't be mirrored? one of those. i think if you search in here discordwise it should pop up. sadly it is late and i am very sleepy from trying to see the aurora 21:40:04 <06mumra> Hmm, mirroring sounds like a plausible problem 21:43:18 03gammafunk02 07* 0.33-a0-228-g06a421242d: Add the Abjuration spell to clua monster info 10(9 minutes ago, 1 file, 11+ 1-) 13https://github.com/crawl/crawl/commit/06a421242d9f 21:48:47 03mumra02 07* 0.33-a0-229-gc8b311e6f6: Remove some dud param comments and fix typo 10(4 months ago, 1 file, 1+ 5-) 13https://github.com/crawl/crawl/commit/c8b311e6f6ea 21:48:47 03mumra02 07* 0.33-a0-230-g41fc793b43: Remove a redundant XXX comment 10(4 months ago, 1 file, 0+ 4-) 13https://github.com/crawl/crawl/commit/41fc793b4387 21:48:47 03mumra02 07* 0.33-a0-231-ge325cd85d1: Remove and clarify a TODO around OOD monsters 10(4 months ago, 1 file, 2+ 1-) 13https://github.com/crawl/crawl/commit/e325cd85d13e 21:48:47 03mumra02 07* 0.33-a0-232-gde153c73cf: Correctly describe Canine Familiar cooldown 10(5 months ago, 1 file, 2+ 1-) 13https://github.com/crawl/crawl/commit/de153c73cfa9 21:48:47 03mumra02 07* 0.33-a0-233-g6fe43b3371: Fix Sniper showing 1% hit chance and bypass shield 10(2 weeks ago, 2 files, 9+ 11-) 13https://github.com/crawl/crawl/commit/6fe43b33715b 21:48:47 03mumra02 07* 0.33-a0-234-g2cf4ae63b3: Comment cleanup: an already-done constify todo 10(3 days ago, 1 file, 0+ 1-) 13https://github.com/crawl/crawl/commit/2cf4ae63b3a9 21:48:47 03mumra02 07* 0.33-a0-235-g6ce201e34c: Clean up an unused function call 10(6 days ago, 1 file, 0+ 3-) 13https://github.com/crawl/crawl/commit/6ce201e34ce0 21:48:47 03mumra02 07* 0.33-a0-236-g2156aa9015: Fix randomness in shield block to-hit calculation 10(15 minutes ago, 4 files, 42+ 17-) 13https://github.com/crawl/crawl/commit/2156aa90153d 21:48:55 <08nicolae> i remember there was some internally-asymmetrical vault i was testing and not all the possible reflections/rotations showed up, i think it was mirroring but yeah, i would have to check 22:35:22 Unstable branch on crawl.develz.org updated to: 0.33-a0-236-g2156aa9015 (34) 22:58:29 Windows builds of master branch on crawl.develz.org updated to: 0.33-a0-236-g2156aa9015 23:30:05 Unstable branch on cbro.berotato.org updated to: 0.33-a0-236-g2156aa9015 (34) 23:49:08 <04melvinkitnick> So, systemd timer that triggers trunk builds service was not triggering, because associated service was "stuck" in running state; apparently chown process on "Installing game" step: bash $ sudo journalctl -u crawl-update-trunk.service ... Oct 04 01:20:59 rex sudo[365928]: crawl-dev : PWD=/var/local/crawl-dev/dgamelaunch-config/crawl-build/crawl-git-repository/crawl-ref ; USER=root ; 23:49:09 COMMAND=/var/local/crawl/DGL/sbin/install-trunk.sh bb4998b30b 0.33-a0-203-gbb4998b30b 34 0 Oct 04 01:20:59 rex sudo[365928]: pam_unix(sudo:session): session opened for user root(uid=0) by (uid=1003) Oct 04 01:20:59 rex dgl[365929]: Installing game Oct 11 08:30:28 rex sudo[365928]: pam_unix(sudo:session): session closed for user root Oct 11 08:30:28 rex dgl[345937]: Terminated Oct 11 08:30:28 rex systemd[1]: crawl-update-trunk.service: Main process 23:49:09 exited, code=exited, status=1/FAILURE Oct 11 08:31:58 rex systemd[1]: crawl-update-trunk.service: State 'final-sigterm' timed out. Killing. Oct 11 08:31:58 rex systemd[1]: crawl-update-trunk.service: Killing process 365953 (chown) with signal SIGKILL. Oct 11 08:32:34 rex systemd[1]: crawl-update-trunk.service: Sending signal SIGTERM to process 365953 (chown) on client request. Oct 11 08:33:28 rex systemd[1]: crawl-update-trunk.service: Processes still 23:49:10 around after final SIGKILL. Entering failed mode. Oct 11 08:33:28 rex systemd[1]: crawl-update-trunk.service: Failed with result 'exit-code'. Oct 11 08:33:28 rex systemd[1]: crawl-update-trunk.service: Unit process 365953 (chown) remains running after unit stopped. Oct 11 08:33:28 rex systemd[1]: Stopped crawl-update-trunk.service - Rebuild crawl (trunk). Oct 11 08:33:28 rex systemd[1]: crawl-update-trunk.service: Consumed 22min 10.657s CPU time. 23:52:06 <04melvinkitnick> bash $ sudo ps -up 365953 USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND root 365953 0.0 0.1 19552 18624 ? D Oct04 0:53 chown -R crawl:crawl /var/local/crawl/DGL/dgldir Anyways. There's still open files listed through lsof, but none belonging to root still in this folder. I'll check this later if needed. A rebuild is currently running. 23:53:38 <04melvinkitnick> Note that I monitor systemd units failed state, though it was not the case here. I'll see how to monitor non-running timers as well. 23:55:17 Monster database of master branch on crawl.develz.org updated to: 0.33-a0-236-g2156aa9015