00:27:35 Unstable branch on cbro.berotato.org updated to: 0.35-a0-32-g2ad184ce96 (34) 00:56:41 Monster database of master branch on crawl.develz.org updated to: 0.35-a0-32-g2ad184ce96 04:37:07 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-5261-gd9800d219b 05:26:33 Unstable branch on crawl.akrasiac.org updated to: 0.35-a0-32-g2ad184c (34) 08:00:28 New branch created: pull/5057 (1 commit) 13https://github.com/crawl/crawl/pull/5057 08:00:29 03CrawlOdds02 07https://github.com/crawl/crawl/pull/5057 * 0.35-a0-33-g3dae1419ed: Don't summon trees under monsters who can't live in them 10(58 minutes ago, 1 file, 4+ 2-) 13https://github.com/crawl/crawl/commit/3dae1419ed8e 18:11:52 03WizardIke02 07* 0.35-a0-33-g7ecffdf5e8: Fix a crash when sealing a mutli square door 10(49 minutes ago, 1 file, 169+ 127-) 13https://github.com/crawl/crawl/commit/7ecffdf5e896 18:22:38 03WizardIke02 07[stone_soup-0.34] * 0.34-b1-33-g45a268f7a6: Fix a crash when sealing a mutli square door 10(60 minutes ago, 1 file, 169+ 127-) 13https://github.com/crawl/crawl/commit/45a268f7a6e0 18:28:10 04Build failed for 08master @ 7ecffdf5 06https://github.com/crawl/crawl/actions/runs/21573917443 18:43:40 arena crash with FULLDEBUG "miscasts 5 pandemonium lord v 20 20-headed hydra delay:0 t:10" 19:22:21 astronautguy (L20 FeMo) ASSERT(defender() && defender()->is_monster()) in 'fineff.cc' at line 1423 failed. (Vaults:2) 19:25:13 <08o​____0> !crash 19:25:19 <04C​erebot> 22590. astronautguy, XL20 FeMo, T:46697 (milestone): https://crawl.akrasiac.org/rawdata/astronautguy/crash-astronautguy-20260202-022220.txt 19:50:20 <09g​ammafunk> notable that only that clang build failed make test 19:51:25 <09g​ammafunk> the two other gcc full debug builds passed make test 19:51:29 <09g​ammafunk> so not sure what to make of that 19:51:42 <09g​ammafunk> I suppose it can just be recreated locally (one would hope) 19:56:35 let's see what happens locally (I use clang18 on ubuntu) 20:02:11 <08o​____0> Not sure why it's crashing exactly but apparently an enemy frostheart's creeping frost can sometimes hurt other enemies? I can sometimes get it to happen 20:03:57 <09g​ammafunk> also, wow, the first thing to stop qw in its tracks in trunk wasn't inventory clua api changes or wall monsters or warding or tesseracts 20:04:17 <09g​ammafunk> it was that lua 5.4 changes the syntax for variable function arguments... 20:04:33 <09g​ammafunk> probably not changed actually in 5.4 but between 5.1 and 5.4 it's changed 20:04:40 <09g​ammafunk> wonder if we used those anywhere in dlua... 20:09:33 <09g​ammafunk> ah, I see, arg.n was actually deprecated in lua 5.1 and available as a compatibility feature 20:09:56 <09g​ammafunk> and it seems crawl has never used it 20:10:25 <09g​ammafunk> wonder how I got the idea to use it for qw, then, maybe elliptic was using it already 20:42:50 hrm, I still have the problem that cerebov vs. test spawner hangs after completion and the test eventually times out and fails 20:48:51 and the miscasts test is doing the same, which makes me think util/fake_pty is b0rked on my system for some reason 21:00:22 <04d​racoomega> Er, I don't think I see any evidence of that happening in the crash log? That Mara is a bound soul on the player's side (and the problem seems to be something to do with the vampiric thrall prop interacting weirdly with Bind Soul. Are you sure you've seen creeping frost hurting the caster's allies? 21:03:09 okay, yeh, that test timed out but didn't otherwise crash here. wonder if it's reproducible in CI 21:20:46 <08o​____0> ⁨ The ironbound frostheart calls freezing cold from the walls. The frost grips your crawling flesh cage! Your crawling flesh cage resists. The frost grips you. You resist. You are encased in ice. The frost grips your crawling flesh cage! Your crawling flesh cage resists. The frost grips Mara! Mara resists. Mara rises to serve you! ⁩ From the crash log ^ Screenshot is from my game with no bound mara 21:20:46 <08o​____0> https://cdn.discordapp.com/attachments/747522859361894521/1467736477717893202/vSki9tv.png?ex=6981779d&is=6980261d&hm=5fce40adab0ca379ef9fed9fe91e868bc45c8c0d8056a328a6e9f32f277c2503& 21:21:54 <08o​____0> I don't think it's a critical part of the crash, just something strange I noticed 21:23:41 <08o​____0> (I don't think the frostheart friendly fire is needed for the crash, mara coming back as a bound soul seems to be a part of it) 21:24:41 <04d​racoomega> ⁨Mara rises to serve you!⁩ isn't implying he was originally hostile here (paradoxically). It's the thrall kill flag not gettin cleaned when they were first bound 21:24:47 <04d​racoomega> If you look up, you can see Mara fighting on their side in previous battles 21:24:56 <04d​racoomega> Wait, what? 21:24:57 <08o​____0> ah ok 21:25:05 <08o​____0> But my screenshot doesn't have a bound mara 21:25:13 <04d​racoomega> That's just a normal Mara?? 21:25:19 <08o​____0> https://cdn.discordapp.com/attachments/747522859361894521/1467737624402854077/ucaOU5G.png?ex=698178ae&is=6980272e&hm=ae9169f6809f58000976d6f79dd1d57bbcb1b190778bcb82d888beefd2372253& 21:25:34 <04d​racoomega> ???? 21:25:40 <08o​____0> oh no I see 21:26:00 <08o​____0> It's the undead mara the log just doesn't say I htink 21:26:08 <08o​____0> Because the living one isn't even next to a wall 21:26:31 <04d​racoomega> How do you have a spectral mara while a real one is alive? (You just made several?) 21:26:40 <08o​____0> Yeah trying to reproduce haha 21:26:53 <08o​____0> yred just makes em when they die 21:26:58 <08o​____0> so just an accident, no bug 21:27:26 <04d​racoomega> Anyway, my theory is that they used Bind Soul, then killed Mara in a way that put the 'turn into a thrall' flag on them. Becoming a bound soul took precedence, but left the flag around for when they next died, and then unhappy things happened when it tried to turn them into a thrall 21:28:14 <04d​racoomega> (Incidentally, I built on clang to try the test failures and it's just passing without incident on my computer) 21:29:21 right, it passed here too aside from the hang/timeout 21:40:28 <09g​ammafunk> concerning though that the test can just randomly fail/crash on clang 21:40:53 <09g​ammafunk> guess we'll just have to wait and see if it pops up again, or something 21:41:05 <09g​ammafunk> would be nice actually if we could get those crash logs 21:41:26 <09g​ammafunk> I wonder if we could do some neat CI thing to store those somewhere 21:43:18 it's possible to register post-job actions; the cache action does so 21:44:10 <09g​ammafunk> probably we could store N crash logs (or up to M megabytes of them) somewhere then 21:44:50 <09g​ammafunk> but actually maybe you'd prefer to transfer them somewhere, since I'm not sure if they'd be accessible directly through github, would have to look into that probably 21:50:28 <04d​racoomega> It would definitely be very helpful 21:50:45 save it as an artifact, only if the test step fails (there are step conditions for that) 21:55:05 because there's some magic in how GHA deals with failure, it'd be `if: failure() && steps.test.conclusion == 'failure'` (assuming `id: test` added to the `make test` step) 22:09:42 https://docs.github.com/en/actions/reference/workflows-and-actions/metadata-syntax#runspost is how you add post-job actions 22:09:50 so there's your magic 22:11:14 hm, looks like that requires defining a custom action 22:11:28 (`actions/cache` being such) 22:13:01 which probably means you can't access the status of the `test` step unless you make the test a local composite action. but that might be worth it in order to get the logs and maybe core file as artifacts 22:14:55 wait, this doesn't need a post action, just have the last step in the test job use that `if` 22:15:06 or even the step following the test 22:15:55 (I've fought a lot with GHA, and it's still confusing) 23:35:41 Unstable branch on crawl.develz.org updated to: 0.35-a0-33-g7ecffdf5e8 (34)