03:34:45 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-5208-geafff8c3b6 04:00:48 fr: black mark working with ranged attacks (anti frustration feature for ranged ds) 04:04:35 Right now it seems to be nekomata 04:06:57 I met him in pan once 04:08:28 now I only need to meet the new holy monster so I can post a picture on it on instragram 04:08:33 with my completed pok�dex 04:08:59 Fr: monster encyclopedia with useless trivia about every monster 04:09:21 <02Monkooky> late uniques are never big killers- people don't see them often enough for them to be a major cause of death 04:13:53 Nekomata were turning up on the orbrun and marmalising people then, but I think that may have been turned off 04:14:10 fr: hand gun mode. You win the game once you find a hand gun and a 8 bit rendition of "hard times come again no more" starts playing 04:14:22 let the credits roll... 07:59:48 (or "people of the sun") 08:27:04 monokles (L14 GrEE) ASSERT(mon.alive() || !god_conducts) in 'spl-damage.cc' at line 560 failed. (Orc:2) 08:34:14 !crashlog monokles 08:34:14 1. monokles, XL14 GrEE, T:16836 (milestone): https://underhound.eu/crawl/morgue/monokles/crash-monokles-20241017-152703.txt 08:35:40 mm, orc sorcerer surrendered and then got crushed by a boulder. did the game try to have it convert after it was dead? 08:36:36 (of late I've been wishing the game made save backups automatically on online crashes, probably with a short expiration like a week) 08:50:49 <06mumra> in saves do we capture the full state of the RNG? so if i played out the exact same set of moves i would get exactly the same result 08:51:18 we're supposed to, yes 08:51:56 iirc that was part of the rearrangement to support seeds and such, to allow more complete reproductions of games 08:53:33 <06mumra> well there are different RNGs you see. so the state of levelgen RNG is captured, but i'm not sure about the RNG used for gameplay like combat, monster movement etc 08:54:01 <06mumra> i'm just thinking if all the user input was captured since the last save, as well as the RNG state, you could replay all the conditions that led to a crash 08:57:29 <06mumra> cos sometimes even if you had a save, reproducing the exact circumstances might be extremely hard 09:01:22 <06mumra> i guess we already have inputs from ttyrecs? 09:02:49 <06mumra> however, i have an idea what went on with this specific crash 09:02:56 <06mumra> because i was just looking at some vaguely related code 09:04:30 <06mumra> in cases like this, the game temporarily sort-of kills the monster, and then properly revives them later in a fineff 09:05:25 <06mumra> so i think the boulder has moved and applied damage to the orc during this window when they're not properly alive 09:16:17 <06mumra> so ... during that assert, the orc should have had flag MF_PENDING_REVIVAL, to indicate they're only mostly dead 09:18:44 Dwargi (L12 DsFi) Crash caused by signal #6: Aborted (Lair:2) 09:19:51 …I never know what to do with those 09:24:00 <06mumra> this ASSERT(mon.alive()) could also have a check for MF_PENDING_REVIVAL, and leave them alone if they have this flag, but it begs the question is anything else going to get tripped up by this weird state? 09:24:49 <06mumra> and why not just revive them immediately and then rather have beogh just shield them from any other damage for the rest of the round? 09:28:49 you'll have to ask DΩ about that one 09:42:44 <06mumra> yeah i don't know enough to implement the right fix 09:44:03 <06mumra> i strongly suspect there will be numerous other ways a similar crash can happen 09:44:13 <06mumra> there are a LOT of asserts that check alive() 09:45:42 <06mumra> it's just relatively rare to have a deathbed surrender and then something that asserts alive() in the same turn 09:46:38 <06mumra> but i don't think you can simply check MF_PENDING_REVIVAL inside alive() because that 10:02:13 <02Monkooky> is there a canonical 'monster allies' iconography in mon spell icons? 10:05:13 <06mumra> Hoarfrost cannons are cannonical 10:11:20 <06mumra> (Jokes aside, this led me to notice there are two hoarfrost_cannonade icons, one under /ice and one under /alchemy ... i assume the latter is meant to be moved to UNUSED) 10:15:48 <06mumra> And also to actually answer the question, seems like they all just have pentagram icon in the corner 10:15:50 <06mumra> https://cdn.discordapp.com/attachments/747522859361894521/1296522068287950959/image.png?ex=671297c5&is=67114645&hm=c91502c8c0f455f6e704f608d164563470323a864e5472fc91dd2fa04c9d7130& 10:18:29 <06mumra> (Except some of them don't:) 10:18:31 <06mumra> https://cdn.discordapp.com/attachments/747522859361894521/1296522744858415195/image.png?ex=67129866&is=671146e6&hm=c5141307d6dd743e0278501817bc54d57e31946c037c9eb2e64f892d6e49274b& 10:18:41 <06mumra> https://cdn.discordapp.com/attachments/747522859361894521/1296522788474978324/image.png?ex=67129871&is=671146f1&hm=c9312cc7fa56d772e127dc270667399e66c3a53eebf576e0e69f8907b0250433& 10:20:39 <06mumra> If you want an easy reference to all the monster summoning spells, summonsdata handily lists them all https://github.com/crawl/crawl/blob/master/crawl-ref/source/spl-summoning.cc#L2177 11:59:45 <02Monkooky> This is in reference to the blink allies spells 12:00:14 <02Monkooky> So, affects allies rather than summons them 12:05:29 <06pleasingfungus> yeah, think pentagrams are summoning 12:05:37 <06pleasingfungus> would consider a heart 12:06:10 <06pleasingfungus> ala the old ‘friendly’ status icon 12:24:04 -!- wantyapps is now known as aplel 12:44:26 -!- aplel is now known as wantyapps 13:02:01 <04dracoomega> I'm not familiar in detail with why revivals are done this way, but a cursory search suggests there was a handful of different bugs that the current approach was intended to fix (even if clearly it has introduced its own new ones): 13:02:04 <04dracoomega> %git df1d555f45ab 13:02:05 <04Cerebot> ebering * 0.28-a0-1064-gdf1d555f45: feat: fineff-ify attitude changing revivals (2 years, 10 months ago, 4 files, 37+ 1-) https://github.com/crawl/crawl/commit/df1d555f45ab 13:02:50 <04dracoomega> (I never directly touched any of this code when overhauling Beogh. Just made it pacify them instead of convert them into followers and adjusted success rates.) 13:03:45 <04dracoomega> The bug with boulder is actually actually slightly more pesky than even that assert might indicate, since the spell doesn't really anticipate situations where the thing it rolled into can be dead and still occupying space 13:04:37 <04dracoomega> (I could make it just not deal damage to a 'dead' enemy, which would avoid this assert, but then it's still arguably buggy in that it will stop dead at a converted enemy instead of pushing anything back) 13:06:20 <04dracoomega> I do think this state of 'not alive, but still occupying space' is definitely something one could anticipate causing unexpected behavior in a number of places. (Though as that commit says, so is changing attitude partway through an effect) 15:38:52 Unstable branch on underhound.eu updated to: 0.33-a0-251-g3dacd37fa5 (34) 17:08:26 <06regret-⸸nde※> (The monster-only summons use a pentagram in the lower-right corner (and bleed out of their frame) mostly because they're placeholders to try and make the majority of monster spells have icons rather than the majority all being paws (and while there's much higher other priorities for the tile artist devs to be getting around to). The player spells all use pentagrams under the specific summon for now, though Forgecraft will 17:08:26 probably need to revise a bunch of spell icons to make a new seperate approach.) 17:12:30 <06regret-⸸nde※> The blink allies away / allies encircling spell icons that have been offered up are... weird, but I don't think they should be super confusable with summons, between different colours and a less emphasized circle. I'll probably edit in hearts to help it out, at any rate. 18:32:44 03Implojin02 07* 0.33-a0-252-gb65ea80194: Fix inhibited regen infoleaking invis mons 10(14 minutes ago, 1 file, 2+ 1-) 13https://github.com/crawl/crawl/commit/b65ea8019427 18:33:35 <03implojin> oops, wrong # for the closed PR 18:36:03 Some of the implications (eg that one might take off SInv before running away) are interesting 19:40:52 <06mumra> Since there's a god involved, it seems reasonable that they'd save their new follower from getting immediately crushed by the boulder that triggered their conversion ... maybe let them swap places with it rather than stopping it dead? It is a weird and horrible special case to have to have ... 19:43:40 weeird and horrible special cases would be very crawl, though 20:21:21 <06pleasingfungus> definitely 20:21:33 <06pleasingfungus> not endorsing the plan, to be clear 20:21:41 <06pleasingfungus> just saying it’s consistent 🙂 22:35:20 Unstable branch on crawl.develz.org updated to: 0.33-a0-252-gb65ea80194 (34) 22:58:00 Windows builds of master branch on crawl.develz.org updated to: 0.33-a0-252-gb65ea80194 23:04:56 03regret-index02 07* 0.33-a0-253-g2bbf05e178: Revive and heavily revise some removed vaults 10(23 minutes ago, 2 files, 263+ 0-) 13https://github.com/crawl/crawl/commit/2bbf05e17846 23:13:23 Unstable branch on cbro.berotato.org updated to: 0.33-a0-253-g2bbf05e178 (34) 23:55:01 Monster database of master branch on crawl.develz.org updated to: 0.33-a0-253-g2bbf05e178