00:37:47 Fork (bcrawl) on crawl.kelbi.org updated to: 0.23-a0-4907-g988a294422 03:22:56 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-4907-g988a294422 05:18:33 03advil02 07* 0.30-a0-973-g497892bb89: fix: tweak some default status file behavior 10(3 days ago, 3 files, 3+ 3-) 13https://github.com/crawl/crawl/commit/497892bb89aa 05:18:33 03advil02 07* 0.30-a0-974-g2ab08093c7: feat: convert webtiles muting to blocking 10(22 hours ago, 3 files, 151+ 79-) 13https://github.com/crawl/crawl/commit/2ab08093c739 05:18:54 Unstable branch on crawl.akrasiac.org updated to: 0.30-a0-972-gf3ab352 (34) 05:20:26 <06advil> those features will need a webtiles restart to be active, I'll rebuild cao and restart it so that some testing can happen 05:23:21 Unstable branch on crawl.akrasiac.org updated to: 0.30-a0-974-g2ab08093c7 (34) 05:25:43 <06advil> one thing I've also considered adding is an allowlist, that would override /ban [all] 05:25:49 <06advil> may be useful, or I may be overthinking? 05:27:36 <06advil> er, /block [all] 05:29:29 Webtiles server stopped. 05:30:06 <06advil> hm, guess I haven't restarted since I fixed that pid bug 05:31:37 Webtiles server started. 05:33:05 Unstable branch on crawl.kelbi.org updated to: 0.30-a0-974-g2ab08093c7 (34) 06:09:00 03advil02 07* 0.30-a0-975-g2ae22b6899: fix: a variable capture bug in block/kick 10(3 minutes ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/2ae22b689988 06:09:05 <06advil> well, that's a bug I'm glad I caught quickly 06:10:48 <06advil> perils of testing locally with only two accounts logged in 06:12:02 Unstable branch on crawl.akrasiac.org updated to: 0.30-a0-975-g2ae22b6899 (34) 06:12:55 Webtiles server stopped. 06:14:22 Webtiles server started. 06:14:41 <06advil> ugh, still something weird with webtiles start/stop 06:18:15 Unstable branch on crawl.kelbi.org updated to: 0.30-a0-975-g2ae22b6899 (34) 06:36:57 <06advil> maybe I'm just confused about what it's supposed to do; the dgl-config script sends ABRT and the server doesn't gracefully handle that. I have a memory that init.d/webtiles should try to shut it down cleanly, but maybe I'm wrong? 06:37:21 <06advil> also possible I changed the script in place on cao and dgl-config overwrote it at some point 06:37:55 <06advil> would /etc/init.d scripts generally be expected to shut down cleanly, or by any means necessary? 06:43:50 <06advil> I wonder if there is a py2-3 change here to do with https://peps.python.org/pep-0475/ that I am not fully grasping 07:18:07 <09gammafunk> @advil I can't add very much other than pointing out that webtiles restart doesn't seem to work, so I always have to webtiles stop then webtiles start. 07:18:58 <09gammafunk> @advil it's also funny that LSB seems to mandate what actions init scripts should support but don't mandate much in the way of what those actions do https://refspecs.linuxbase.org/LSB_3.0.0/LSB-PDA/LSB-PDA/iniscrptact.html 07:19:58 <09gammafunk> and I guess LSB isn't really even used any more 07:21:49 <09gammafunk> and I guess we can't rely on something like a debian standard since it's all based on systemd right now. To be honest, I'd like to have an official systemd script for starting/stopping webtiles as well 07:29:32 -!- halv1 is now known as halv 07:35:40 <06advil> yeah, init.d is supposed to be deprecated 07:35:50 <06advil> writing a systemd script is probably pretty easy 07:37:09 <06advil> I found https://github.com/crawl/dgamelaunch-config/commit/05b177621e228dc36af608dc91948f136f2e3f64 from 2013 which just says ABRT is faster, but I'm pretty sure the current handling of TERM is a lot better than it was 07:37:40 <06advil> I guess thinking about it more, probably SIGABRT shouldn't aim to gracefully exit, but the stop script also shouldn't send it 07:39:11 <06advil> I haven't tried it in a while, what does it do? just "failed to restart"? 07:40:44 <09gammafunk> It does cause webtiles process to end, but doesn't successfully restart it. No output error, just: > Restarting webtiles (via systemctl): webtiles.service. 07:40:50 init.d isn't _very_ deprecated, but also if your init.d script has an LSB-style BEGIN INIT INFO block it will keep working. 07:41:05 <09gammafunk> interestingly 07:41:19 <09gammafunk> it doesn't seem to leave webtiles in a state where it starts properly via webtiles start 07:41:43 <09gammafunk> I have to webtiles stop and then webtiles start after running webtiles restart before webtiles service is working 07:42:16 <06advil> heh we don't even suupport all the required commands 07:42:28 <09gammafunk> heh, right 07:43:20 <09gammafunk> We should probably make a systemd script that's official and recommended and maybe just leave a cleaned up init script in the repo for those who want to use init for whatever reason 07:55:06 <06advil> it looks like the described behavior of killproc in the lsb scripts is way more right than what we're currently doing 07:55:47 <06advil> though tbh probably no one has ever messed with this beside neil because it's a huge pain to change/debug this sort of script in a non-disruptive way 07:57:46 <06advil> unrelated: I was recently perplexed to see a similar pattern of slow IO on cbr2, though with the alerts being more rare and cao's recent thing of occasionaly many-second blocking not happening 07:58:07 <06advil> which does suggest that something about the webtiles server, or its behavior in combination with things I don't really understand about kernel-level IO, are causing it 07:58:32 <06advil> there's a solution, which is to put IO in a separate thread, and this is roughly standard in asyncio apps 07:59:03 <06advil> but I still don't understand what would cause blocking on IO that can be in the 100s of ms, or even how to approach debugging this 07:59:33 <06advil> I've partly chalked this up to cao's old disk, but I think that's wrong; iirc (not 100% sure) CBR2 even has an ssd, not an hdd 08:00:32 <06advil> (it's even making me wonder if there's something wrong with my code to detect these issues, but I'm pretty sure it's ok) 08:02:39 <06advil> maybe my expectations about code that is doing a fair amount of disk access are just wildly off? noticeable cache penalties (or sth) are just inevitable? 08:03:44 <06advil> I should probably rig up some sort of setup where I have a bunch of qws playing at once locally and see if I can replicate 08:06:37 <09gammafunk> nice. I guess you might want to use the tagged release of qw and 0.29 stable in order to not have qw itself be problematic. qw trunk is much less tested (even if I haven't merged my big branch yet), and there may be breakage from 0.30 I don't know about yet 08:06:51 <06advil> example, here's the last some hours of slow io errors from CBR2 (sry irc): 2023-03-31 00:56:56,678 WARN: Slow IO: write '/crawl-master/webserver/run/status': 0.709s 2023-03-31 03:00:26,526 WARN: Slow IO: write '/crawl-master/webserver/run/status': 2.060s 2023-03-31 03:17:30,488 WARN: Slow IO: write_ttyrec_chunk '/dgldir/ttyrec/Coron/2023-03-31.03:46:05.ttyrec': 0.434s 2023-03-31 03:18:22,416 WARN: Slow IO: write 08:06:51 '/crawl-master/webserver/run/status': 0.280s 2023-03-31 03:20:28,243 WARN: Slow IO: write '/crawl-master/webserver/run/status': 0.728s 2023-03-31 04:01:56,106 WARN: Slow IO: write_ttyrec_chunk '/dgldir/ttyrec/Newt/2023-03-31.07:55:04.ttyrec': 0.359s 2023-03-31 08:41:54,856 WARN: Slow IO: write_ttyrec_chunk '/dgldir/ttyrec/jasper/2023-03-31.12:26:40.ttyrec': 0.491s 2023-03-31 10:36:55,652 WARN: Slow IO: write_ttyrec_chunk 08:06:52 '/dgldir/ttyrec/Brigand/2023-03-31.14:35:21.ttyrec': 0.566s 08:07:17 <06advil> it happens a lot less than on cao, but when it happens, the magnitude of the blocking is really similar, which was just completely unexpected to me 08:07:49 <06advil> for status I do think that the current settings are trying to write much too often, maybe there's something where it trips over itself with many successive open/closes 08:07:59 <06advil> I don't think that can explain the ttyrec penalties though 08:23:21 <03robertxgray> I'm not familiar with the webserver yet, but note the main repo provides a sample systemd script already: https://github.com/crawl/crawl/blob/master/crawl-ref/source/webserver/contrib/dcss-webtiles.service 08:24:05 <06advil> ah good point 08:24:16 <06advil> that won't work as-is for dgl-config setups, but it's probably pretty close 08:24:48 <06advil> (also, welcome to being an orange name!) 08:25:47 <03robertxgray> thanks 😃 09:02:11 <09gammafunk> oh, right! alex wrote that years ago 10:16:36 03PleasingFungus02 07* 0.30-a0-976-g8c80f4669f: Underscore a fix (nicolae) 10(62 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/8c80f4669fd9 10:25:31 -!- bobby-tables is now known as wpa 10:34:29 Unstable branch on crawl.kelbi.org updated to: 0.30-a0-976-g8c80f4669f (34) 10:55:36 03kate-02 07* 0.30-a0-977-gc60b083988: Fix a killer klown message (#3060) 10(38 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/c60b08398820 10:55:36 03kate-02 07* 0.30-a0-978-ga022fb42d2: Fix a targeting check for Plasma Beam (#3057) 10(13 minutes ago, 1 file, 12+ 2-) 13https://github.com/crawl/crawl/commit/a022fb42d237 11:07:18 Unstable branch on crawl.kelbi.org updated to: 0.30-a0-978-ga022fb42d2 (34) 11:29:46 <10PleasingFungus> @kate acrobat suggested that plasma beam might be simpler if it fired one beam that did split fire/elec damage 11:31:03 <12ebering> Think of the bolt bouncing implications (unless bolt bouncing of plasma beam was removed) 11:31:10 <10PleasingFungus> yeah it doesn’t bounce 11:31:15 <05kate> that seems like it'd probably be a bunch more complicated on the code side, and not much simpler on the player side maybe? 11:31:25 <10PleasingFungus> removing bolt bouncing was a core plasma beam design goal 🙂 11:31:53 <10PleasingFungus> i’m not sure it’s be more complex in code? cf the bug you just fixed, for example 11:32:00 <10PleasingFungus> but i don’t feel strongly about it 11:33:20 <10PleasingFungus> just floating the idea 11:33:20 <12ebering> Sad 11:33:20 <10PleasingFungus> i specifically hoped it’d make spellforged servitor support simpler 11:33:20 <10PleasingFungus> (i have a WIP for that - it’s a pain…) 11:37:44 void0 (L6 GnIE) ERROR in 'tileweb.cc' at line 234: Socket write error: Resource temporarily unavailable (D:4) 11:37:44 <05kate> i guess i'm just thinking of how partially-resistible beams work currently, making that support beams that are resistible but in two different ways sounds unfun. also no strong opinion either way though, i don't think it's a problem or particularly complicated as-is on the player side at least 11:40:46 -!- 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 11:40:46 -!- 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 12:08:27 @PF: In trunk webtiles builds, trying to lookup the description of Dith shows "Buggy UI of type describe-god", both through ?G and ^, I bisected this to https://github.com/crawl/crawl/commit/b9aa53057d83ec83a46ad108ffec7826a5d27d99 , seems like the piety display change broke something 12:09:06 <10PleasingFungus> oh, i saw people reporting that and was planning to look into it 12:09:10 <10PleasingFungus> seems like a webtiles specific thing 12:09:14 <10PleasingFungus> thanks for the bisect! 12:09:17 yes, it doesn't occur in local tiles builds 12:24:40 03PleasingFungus02 07* 0.30-a0-979-gadcbbbcd72: Initialize explore mode in webtiles 10(57 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/adcbbbcd723e 12:29:55 03PleasingFungus02 07* 0.30-a0-980-g89a915e538: Welcome Implojin to the devteam! 10(43 seconds ago, 2 files, 2+ 1-) 13https://github.com/crawl/crawl/commit/89a915e538d7 12:35:45 Unstable branch on crawl.kelbi.org updated to: 0.30-a0-980-g89a915e538 (34) 12:35:55 <12ebering> 🙌 12:36:00 <12ebering> Welcome Friends 12:36:53 <12ebering> @Implojin @robertxgray embrace the oranging 12:38:32 <08Implojin> hi hi 12:59:18 03PleasingFungus02 07* 0.30-a0-981-ga366339e96: Remove gauntlet shapeshifters (CarefulOdds) 10(63 seconds ago, 1 file, 0+ 7-) 13https://github.com/crawl/crawl/commit/a366339e96ce 13:05:13 <09hellmonk> Hello, 13:05:38 Unstable branch on crawl.kelbi.org updated to: 0.30-a0-981-ga366339e96 (34) 13:15:06 03PleasingFungus02 07* 0.30-a0-982-g745981c80b: Fix Dith webtiles UI 10(3 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/745981c80b48 13:15:06 03PleasingFungus02 07* 0.30-a0-983-g7e3032bf44: Log errors with buggy UIs 10(50 seconds ago, 1 file, 1+ 0-) 13https://github.com/crawl/crawl/commit/7e3032bf4487 13:17:24 <06advil> can also just rethrow when debugging locally 13:18:13 <10PleasingFungus> i figured this would be useful for users who want to report issues 13:18:48 <06advil> yeah, fair 13:21:05 Unstable branch on crawl.kelbi.org updated to: 0.30-a0-983-g7e3032bf44 (34) 15:24:13 Unstable branch on underhound.eu updated to: 0.30-a0-983-g7e3032bf44 (34) 15:56:30 <06advil> based on https://github.com/crawl/crawl/issues/3064 I suspect that rampage issue I fixed by revert probably happened on going downstairs 15:59:47 <08Implojin> i looked at that briefly the other day and got the impression it might have been a player-relative move delta being passed in as a level-relative coord def, but i didn't have time to track it down 16:01:06 <10PleasingFungus> the code for that was pretty scary 16:01:10 <08Implojin> yes 16:01:16 <08Implojin> i hate everything about the way rampage is implemented 16:01:17 <10PleasingFungus> geekosaur suggested we shouldn't be firing beams while rendering, and i think i agree 16:01:18 <10PleasingFungus> lol 16:01:24 <10PleasingFungus> idk it's not that bad 16:01:43 <10PleasingFungus> but if we bring indicators, i think we should calculate and cache em at the start of the player's turn 16:01:49 <10PleasingFungus> rather than at render time 16:01:55 <10PleasingFungus> *bring rampage indicators back 16:02:03 <08Implojin> agree re caching 16:02:15 <13Sastreii> i made something for april fools 16:02:16 <13Sastreii> https://cdn.discordapp.com/attachments/747522859361894521/1091497720054829087/Schermata_2023-04-01_alle_00.58.05.png 16:02:20 <13Sastreii> https://cdn.discordapp.com/attachments/747522859361894521/1091497741735174285/shrek.png 16:02:22 <08Implojin> didn't make sense to be firing bolts as the tiles are rendered 16:02:52 <06advil> oh, maybe; I was thinking the player is actually at 0,0 at some point during a level transition 16:03:01 <08Implojin> hmmm 16:03:02 <08Implojin> also possible 16:03:40 <10PleasingFungus> lmao 16:03:42 <10PleasingFungus> get out of my swamp! 16:04:30 <10PleasingFungus> important question: in this scenario, who is Donkey 16:05:12 <08nicolae> the player 16:05:25 <10PleasingFungus> wow 16:05:32 <10PleasingFungus> also, thanks for the fix! 16:05:45 03nicolae02 07* 0.30-a0-984-g4aa63f3706: Fix compressed_dpeg_circular_temple again (Sentei) 10(89 seconds ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/4aa63f3706d3 16:06:18 <08nicolae> there were 16 H glyphs and 19 I glyphs, once i counted those the problem and solution presented themselves 16:06:25 <03robertxgray> not fair, I want those ogres, I will pay for the copyright complaints 16:07:02 <13Sastreii> mmm... actually i was more busy thinking who is the perfect Lord Faarquad, which is the best character ever created without any doubt 16:07:15 <10PleasingFungus> lmao 16:07:20 <10PleasingFungus> (feels like Donkey should be maybe a hound, or Natasha? but neither of those seem like amazing fits) 16:07:23 <10PleasingFungus> lord faarquad... hm 16:07:43 <10PleasingFungus> maybe an early unique, a sigmund or terence or edmund 16:07:44 <13Sastreii> ijib? 16:07:52 <10PleasingFungus> oh, ijyb is short 16:07:52 <10PleasingFungus> wait 16:07:52 <10PleasingFungus> robin 16:07:59 <10PleasingFungus> absolutely 100% gotta be robin, right? 16:08:10 <13Sastreii> robin fits very well 16:08:20 <10PleasingFungus> who else is sending all their minions to certain death 16:08:37 <10PleasingFungus> those ogres you will get 16:10:01 <13Sastreii> i'm amused by the fact that i successfully brought chaos with shrek in a serious dev convo 16:12:38 <10PleasingFungus> you gotta have priorities. 16:18:22 03PleasingFungus02 07* 0.30-a0-985-g0638232e92: Fix: minor issue with ogre tile colors (Sastreii) 10(63 seconds ago, 6 files, 20+ 0-) 13https://github.com/crawl/crawl/commit/0638232e923c 16:18:36 Unstable branch on crawl.kelbi.org updated to: 0.30-a0-984-g4aa63f3706 (34) 16:18:37 <10PleasingFungus> ^ this is called 'being sneaky' 16:19:12 <13Sastreii> yeah i was about to say, that's the perfect commit title 16:19:42 <10PleasingFungus> (q) should this actually happen every year (a) probably not, but it will be very funny when i forget and then get surprised next year 16:21:10 <05kate> joke commits, the one way to make PF write a serious commit title... 16:22:11 <13Sastreii> lmfao 16:26:47 <05kate> (also i'm pro leaving it in for next year so that it can take everyone by surprise again) 16:53:51 Unstable branch on crawl.kelbi.org updated to: 0.30-a0-985-g0638232e92 (34) 17:02:29 <09gammafunk> wow, a burn so bad it should be called hellfire 17:05:54 <10PleasingFungus> the other advantage of leaving the tile around is to let goofballs use it as their player tile 17:06:03 <09gammafunk> oh yeah 17:06:10 <09gammafunk> will have to add that to randomtiles 17:06:43 <09gammafunk> which reminds me, what are our new/removed monsters for 0.30 so far? skysharks I know about 17:06:51 <09gammafunk> ??changelog 17:06:52 <04Cerebot> changelog[1/3]: https://github.com/crawl/crawl/blob/master/crawl-ref/docs/changelog.txt - often incomplete and/or out of date, see {changelog[3]} for an exhaustive list of changes. 17:07:27 <09gammafunk> probably that's it, don't even see skysharks in the changelog quite yet 17:31:39 <09hellmonk> some 17:31:46 <09hellmonk> BODY 17:32:24 <12ebering> man haven't played crawl in a while and going to a conference so gonna be super busy but really feel the need to play an OgRe 17:42:58 <09gammafunk> still hard to believe that we brought one of the most epic combo words back 21:52:17 03PleasingFungus02 07* 0.30-a0-986-g7f2e898f52: Update the changelog 10(18 seconds ago, 1 file, 59+ 15-) 13https://github.com/crawl/crawl/commit/7f2e898f5294 21:57:40 <10PleasingFungus> https://cdn.discordapp.com/attachments/747522859361894521/1091587162740367500/image.png 22:03:40 Unstable branch on crawl.kelbi.org updated to: 0.30-a0-986-g7f2e898f52 (34) 22:05:25 <13Sastreii> Ah yes, chaos and confusion starting to spread like a wildfire😈 22:30:22 <09gammafunk> dammit, sleepcaps too, forgot about those 22:34:57 Unstable branch on crawl.develz.org updated to: 0.30-a0-986-g7f2e898f52 (34) 22:57:00 Windows builds of master branch on crawl.develz.org updated to: 0.30-a0-986-g7f2e898f52 22:57:52 <10PleasingFungus> http://crawl.develz.org/wordpress/trunk-updates-1-april-2023-and-tournament-announcement assume there are a bunch of typos and stuff but i'm going to sleep 23:37:48 Unstable branch on cbro.berotato.org updated to: 0.30-a0-986-g7f2e898f52 (34) 23:54:25 Monster database of master branch on crawl.develz.org updated to: 0.30-a0-986-g7f2e898f52