01:01:24 03CrawlOdds02 07* 0.35-a0-721-g5c2ea9cce7: Fix exclusions going beyond walls on reload 10(8 hours ago, 1 file, 12+ 11-) 13https://github.com/crawl/crawl/commit/5c2ea9cce77b 01:11:01 03CrawlOdds02 07* 0.35-a0-722-ge8d8efbb73: Make autoexplore take visited transporters 10(10 hours ago, 1 file, 118+ 80-) 13https://github.com/crawl/crawl/commit/e8d8efbb7346 01:47:48 <11Odds> Pain bond is broken in a funny way where the damage can blow up via recursively calling it. It does seem to deliberately recurse, but also that seems to be a terribly bad idea... 01:47:49 <11Odds> https://cdn.discordapp.com/attachments/747522859361894521/1528322453074350101/image.png?ex=6a5de0b3&is=6a5c8f33&hm=1fbc271a9a6b354bc81bca69560db54c81c0fecbce301a1c22f794946c333b4e& 01:54:14 <11Odds> I reckon we should just make each monster take damage from a pain bond event only once, and give it a little numbers buff to compensate (this will mean the original target takes no extra damage, which seems correct to me) 01:56:48 <12ge0ff> > that seems to be a terribly bad idea What do you mean? Killing three pan lords from one stab feels like a very good result! :kobold: 01:57:12 <11Odds> The hard part is arranging them in a triangle with pain bond active 🙂 01:58:23 <11Odds> (This is all it takes for the bond to radiate until one of them dies. Actually Cerebov survived as this might stab as the fourth corner of the square) 03:21:12 03WizardIke02 07* 0.35-a0-723-gd0518a0150: Fix wizmode placement of encompass vault generating duplicate uniques 10(27 minutes ago, 1 file, 9+ 13-) 13https://github.com/crawl/crawl/commit/d0518a01503d 03:34:02 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-5261-gd9800d219b 05:06:49 New branch created: pain-bond-fix (1 commit) 13https://github.com/crawl/crawl/tree/pain-bond-fix 05:06:49 03CrawlOdds02 07[pain-bond-fix] * 0.35-a0-724-g335319cf3a: Prevent pain bond dealing infinite damage 10(3 hours ago, 3 files, 40+ 36-) 13https://github.com/crawl/crawl/commit/335319cf3afc 05:08:53 <11Odds> I'd suggest something like that, modulo balance (which is kinda tricksy here because the current pain bond has such weird behaviour) 06:57:27 Don't you DARE NERF PAIN BON—oops I'm too late 06:58:38 wait you buffed the numbers a bit to compensate okay we can maybe still be friends 07:03:20 <04dracoomega> Wow, what? (Mind you, in large and dense groups of monsters, it did tend to end up as 'they all die in the next few hits', anyway, though to some degree that felt like the reward for making it that far) 07:04:36 <04dracoomega> Though I admit I'm unclear on how 1 damage could radiate infinitely. Even at distance 1, doesn't the 1 damage round down to 0? (ie: 1 * 3 / 5)? 07:09:08 <11Odds> No, it randomly rounds. So if you have a square of 4 monsters, the expected damage the first radiate turns out to be (0.6 + 0.6^2 + 0.6^3) > 1. 07:09:40 <11Odds> (You do need a square not a triangle actually - I'd missed the second bug which makes the triangle case 0.96 damage rather than 1.2) 07:10:36 That is a really cool find; how'd you notice it anyway? 07:11:19 <04dracoomega> I am not certain that is infinite in practice, though, regardless of what the average is? Since it only needs to round down on a single pass to come to a screeching halt, no? 07:11:28 <04dracoomega> (This is not at all a counterargue to this being silly, to be clear) 07:12:03 <11Odds> No - the expected damage in the nth "radiation" is (0.6 + 0.6^2 + 0.6^3)^n. Any particular chain certainly terminates, but there are more and more chains as we go on 07:12:35 <11Odds> There's an old issue which didn't get reproduced, and then I read the code 07:12:44 <11Odds> https://github.com/crawl/crawl/issues/3006 07:12:59 <11Odds> (Having a fun "old issues" weekend) 07:14:42 <11Odds> (This is just a branch atm, and the numbers may well be a bit off) 07:16:08 <04dracoomega> If the expected total damage from anything with a bunch of adjacent monsters was actually infinite, why was it hard to reproduce in practice, I wonder? 07:17:06 <11Odds> It does have to be somewhat dense. A triangle and another monster should do it, though. And you often don't notice because one of them dies and breaks the cycle of doom 07:17:25 I think it's easy to pass off as you just killing things. When you get pain bond you're usually using a high-damage weapon anyway (not a poison dart) 07:17:33 (Or doing para stabs, etc) 07:17:53 <11Odds> It only took one attempt to kill three of my four pan/helllords in that square 07:19:25 <04dracoomega> Yeah, I don't have an immediate intuition of what the numbers 'ought' to be either. Though it's worth keeping in mind that - no matter this being a silly thing that makes sense to change - Usk tended not to be considered a high-tier god as it was, and pain bond was deliberately a battle-winning payoff in many cases. I'd probably like to stare at it a bit, later. 07:20:40 (Yeah I don't want to talk about how many Usk games I've lost) 07:21:36 <11Odds> Yeah sounds good. One important bit is not amplifying the damage to the original monster at all, which we could change. 07:22:20 <11Odds> (Another nice thing about not damaging the original monster is that it prevents various pain bond related bugs like that Blorkula one I fixed recently) 07:23:20 I will say that for Usk the most annoying thing is how confusing the cooldowns for stuff like solo and pain bond are and how that's separate from the piety thresholds, so you can hit a piety threshold, be told that Usk's gonna give you a solo, and then nothing happens for 10 turns 08:05:21 <11Odds> @dracoomega re: https://github.com/crawl/crawl/issues/4827 - do you still think as you commented there that it would be nice for coglins to be allowed to swap an offhand item in for either weapon (including if the "offhand" weapon is distortion, cursed, etc)? Seems like it would be good to me (and pretty doable) 08:08:47 <04dracoomega> Yes, for sure. The fact that - in this specific case - they're not interchangeable is just a consequence of the backend and not something I think should have player-facing effects, ideally. 08:09:29 <04dracoomega> I think handling this 'correct' is relatively straightforward, just mildly fiddly to implement (and I never got around to it) 08:10:16 <11Odds> Cool. Yeah I think there's some wrangling to do to choose the item, deal with unequippable things, etc, but nothing too painful. I'll have a look. 08:13:35 <04dracoomega> Fortress crab already does some secret backend moving of weapon between slots; not sure if looking at this might be useful reference or not, at this point. 08:15:42 <11Odds> Thanks! 11:20:31 03CrawlOdds02 07* 0.35-a0-724-g9e9c337617: Let coglins equip offhand items without unequip 10(3 hours ago, 3 files, 93+ 32-) 13https://github.com/crawl/crawl/commit/9e9c3376177c 11:20:31 03CrawlOdds02 07* 0.35-a0-725-g8e159b2936: Fix a spurious coglin distortion removal message 10(3 hours ago, 1 file, 7+ 0-) 13https://github.com/crawl/crawl/commit/8e159b293624 11:20:31 03CrawlOdds02 07* 0.35-a0-726-g2dce783617: Give Ozocubu's armour a trivial targetter 10(2 hours ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/2dce78361746 12:38:32 <12ge0ff> @Odds, re: 049aa0cdb, "Stop useless casts of scorch, permafrost and plasma beam". From the title I was expecting this commit to fix another old bug - casting Scorch vs. rF+++ monsters. Currently, the spell is active on the quiver/spell list even when there are only fire-immune monsters nearby, like crimson imps/fire elementals, but the spell does nothing to them. And it's not an obvious behavior: Cerebov's sword can remove a pip of rF 12:38:32 even from fire-immune monsters, so the player might think that Scorch would do the same. So it would be great if the spell wasn't marked as useful in such situations too. 12:42:11 <11Odds> Noted 12:42:25 <11Odds> I'll fix that one some time 13:47:48 03CrawlOdds02 07* 0.35-a0-727-ga3eea6ddbd: Fix a webtiles menu issue 10(2 hours ago, 1 file, 4+ 1-) 13https://github.com/crawl/crawl/commit/a3eea6ddbd66 13:47:48 03CrawlOdds02 07* 0.35-a0-728-ga8e057999a: Fix !experience not selecting djinni magic skills near to 27 10(73 minutes ago, 1 file, 8+ 0-) 13https://github.com/crawl/crawl/commit/a8e057999a8f 13:47:48 03CrawlOdds02 07* 0.35-a0-729-gf2919c6e55: Fix problem with djinni not training at max skill 10(42 minutes ago, 1 file, 7+ 2-) 13https://github.com/crawl/crawl/commit/f2919c6e558a 15:14:16 03CrawlOdds02 07* 0.35-a0-730-ga04e86051b: Hide tooltips when statuses expire 10(3 minutes ago, 1 file, 5+ 1-) 13https://github.com/crawl/crawl/commit/a04e86051bdf 15:43:09 Unstable branch on underhound.eu updated to: 0.35-a0-729-gf2919c6e55 (34) 22:36:02 Unstable branch on crawl.develz.org updated to: 0.35-a0-730-ga04e86051b (34) 22:59:58 Windows builds of master branch on crawl.develz.org updated to: 0.35-a0-730-ga04e86051b 23:33:24 Unstable branch on cbro.berotato.org updated to: 0.35-a0-730-ga04e86051b (34) 23:56:23 Monster database of master branch on crawl.develz.org updated to: 0.35-a0-730-ga04e86051b