00:23:17 kate-: I'm aware of that contradiction and am not 100% on giving them the spectral club to start, but I wanted to try it out 00:23:26 and also give people a chance to try the ego quickly 01:38:47 Unstable branch on crawl.develz.org updated to: 0.26-a0-365-g11c2b627af (34) 02:20:31 03Aidan Holm02 07* 0.26-a0-366-gd6aadd2: Fix clouds being drawn under monsters/items (Yermak) 10(81 seconds ago, 1 file, 1+ 2-) 13https://github.com/crawl/crawl/commit/d6aadd2fae8a 02:28:05 hey aidanh I saw that you were just in draw code; would you have any idea what went wrong with my (fork's) translucent water? It worked fine on local/windows build but on the server it failed to work like I wanted. (supposed to be a translucent feat drawn over the floor; on local worked fine; on webtiles it either didn't draw the floor under it or 02:28:05 it turned the alpha into black making just DARK water tiles instead of semi-clear ones) 02:31:58 Unstable branch on crawl.kelbi.org updated to: 0.26-a0-366-gd6aadd2fae (34) 02:32:53 i'm not familiar with your fork, but presuming the code structure is basically the same, you should be aware of the difference in render models between local tiles and web tiles 02:33:35 local tiles first accumulates tiles into buffers based (more or less) on the tilesheet, in a particular order, and then renders each buffer in a particular order 02:34:30 whereas webtiles just renders tiles one after the other; there is no implicit or explicit sorting 02:36:30 (do note that while i'm happy to answer questions about mainline crawl's code, with the hope that the information transfers, fork development is still off-topic) 02:38:08 Yea I'm looking at where your last commit was (in cell_renderer) to attempt a fix. I'm guessing I need to change something in draw background around this section." 02:38:38 ...or reorder the enum so the clear water comes after DNGN_FIRST_TRANSPARENT... 03:10:37 !kw day 03:10:37 Keyword: day => end>${now()-interval('1d')} 03:10:37 !kw twodays end>${now()-interval('2d')} 03:10:37 Defined keyword: twodays => end>${now()-interval('2d')} 03:10:37 !kw threedays end>${now()-interval('3d')} 03:10:37 Defined keyword: threedays => end>${now()-interval('3d')} 03:11:52 Unstable branch on crawl.beRotato.org updated to: 0.26-a0-366-gd6aadd2 (34) 03:31:04 Fork (bcrawl) on crawl.kelbi.org updated to: 0.23-a0-3470-gaa44c949d2 03:46:15 Fork (bcadrencrawl) on crawl.kelbi.org updated to: 0.22.1-2863-g4211a4d3ca 04:15:55 -!- Yermak_ is now known as Yermak 08:43:49 New branch created: directn-widget (9 commits) 13https://github.com/crawl/crawl/tree/directn-widget 08:43:50 03Aidan Holm02 07[directn-widget] * 0.26-a0-367-gb9eb94e: Fix widgets having a size of 0x0 by default 10(6 hours ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/b9eb94e09650 08:43:50 03Aidan Holm02 07[directn-widget] * 0.26-a0-368-g7d3f501: Refactor startup code 10(10 months ago, 5 files, 36+ 32-) 13https://github.com/crawl/crawl/commit/7d3f50155626 08:43:50 03Aidan Holm02 07[directn-widget] * 0.26-a0-369-g4f64a06: Wrap direction chooser UI in a widget 10(4 hours ago, 6 files, 125+ 66-) 13https://github.com/crawl/crawl/commit/4f64a06377de 08:43:50 03Aidan Holm02 07[directn-widget] * 0.26-a0-370-gae94ea5: Don't force screen update when printing stats 10(2 hours ago, 10 files, 14+ 11-) 13https://github.com/crawl/crawl/commit/ae94ea54d2d0 08:43:50 03Aidan Holm02 07[directn-widget] * 0.26-a0-371-g1e9b5db: Don't force screen update in viewwindow 10(2 hours ago, 39 files, 90+ 3-) 13https://github.com/crawl/crawl/commit/1e9b5dbab53a 08:43:50 03Aidan Holm02 07[directn-widget] * 0.26-a0-372-g3dc048d: Remove redraw_console_sidebar() hack 10(71 minutes ago, 3 files, 0+ 53-) 13https://github.com/crawl/crawl/commit/3dc048d44a1c 08:43:50 03Aidan Holm02 07[directn-widget] * 0.26-a0-373-g4528f9d: Don't force screen update in redraw_screen 10(36 minutes ago, 49 files, 171+ 10-) 13https://github.com/crawl/crawl/commit/4528f9da3680 08:43:50 03Aidan Holm02 07[directn-widget] * 0.26-a0-374-g0a399d6: Don't update screen when exiting last UI layout 10(34 minutes ago, 1 file, 0+ 3-) 13https://github.com/crawl/crawl/commit/0a399d6a6532 08:43:50 03Aidan Holm02 07[directn-widget] * 0.26-a0-375-g8b0982a: Remove redundant redraws in direction chooser 10(81 minutes ago, 1 file, 2+ 13-) 13https://github.com/crawl/crawl/commit/8b0982a8820f 08:45:58 oh good, is this the direction_chooser reform? 08:48:29 yep 08:49:49 there's some fixes for excessive redraws as well 08:56:10 huh 08:56:19 TiL you can case a redraw of screen from lua 08:57:26 I can't seem to find any use of it though 08:58:02 I guess it's just clua that we make available, but I wonder how wise that is 09:19:31 it's a bad idea and all that should be deprecated and ripped out 09:20:13 well, for the redraw function at least, I don't see any of our clua using it 09:20:21 probably safe to go ahead and do so 09:20:36 and I don't know of any important community clua thing using something like that 09:21:24 qw doesn't use it either 09:38:05 The build passed. (directn-widget - 8b0982a #14006 : Aidan Holm): https://travis-ci.org/crawl/crawl/builds/709689630 09:48:58 I don't think we make any promises of api stability 09:49:02 I pulled out all the food stuff 10:04:54 yeah, we don't really have any big community clua-using projects where we'd worry too much; I guess the only concern might be for admins who have many reports if something breaks, but there's nothing really like that 10:05:04 in terms of very high usage 10:16:34 the only case i can recall is someone mentioning a script to automate running around pillars that used the direction chooser 10:19:56 heh, I recall that, and it was very recent 10:20:16 there were other sort of cute projects like UV4's "red carpet", partially responsible for ebering's trap reform 10:20:46 but I was more musing about where breaking api might actually cause us some significant headaches 10:20:55 the closest thing that comes to mind might be breaking HDA's rcs 10:21:25 those are probably the most widely used lua rc things out there; some of my things are used fairly well but I don't think nearly as much 10:21:43 and even if we somehow broke hdarc, I don't think it would cause us real headaches 10:23:03 just removing those functions is easy enough to revert in any case 11:07:29 Chei still has slow hunger on piety gain message 13https://crawl.develz.org/mantis/view.php?id=12321 by hayenne 11:48:12 03Aidan Holm02 07[directn-widget] * 0.26-a0-368-g34a7423: Extract ui::MouseEvent -> wm_mouse_event conversion 10(19 minutes ago, 3 files, 20+ 13-) 13https://github.com/crawl/crawl/commit/34a74239f36d 11:48:12 03Aidan Holm02 07[directn-widget] * 0.26-a0-369-g5a2f5f8: Reimplement direction chooser UI as a widget 10(7 hours ago, 6 files, 141+ 67-) 13https://github.com/crawl/crawl/commit/5a2f5f8edd2e 11:48:12 03Aidan Holm02 07[directn-widget] * 0.26-a0-370-g9d69a02: Don't force screen update when printing stats 10(5 hours ago, 10 files, 14+ 11-) 13https://github.com/crawl/crawl/commit/9d69a0231071 11:48:12 03Aidan Holm02 07[directn-widget] * 0.26-a0-371-gfc10b71: Don't force screen update in viewwindow 10(5 hours ago, 39 files, 90+ 3-) 13https://github.com/crawl/crawl/commit/fc10b7151e37 11:48:12 03Aidan Holm02 07[directn-widget] * 0.26-a0-372-g20cb03b: Remove redraw_console_sidebar() hack 10(4 hours ago, 3 files, 0+ 53-) 13https://github.com/crawl/crawl/commit/20cb03b44e4f 11:48:12 03Aidan Holm02 07[directn-widget] * 0.26-a0-373-g0c3ae41: Don't force screen update in redraw_screen 10(4 hours ago, 49 files, 171+ 10-) 13https://github.com/crawl/crawl/commit/0c3ae4171de3 11:48:12 03Aidan Holm02 07[directn-widget] * 0.26-a0-374-ga21bf3d: Don't update screen when exiting last UI layout 10(4 hours ago, 1 file, 0+ 3-) 13https://github.com/crawl/crawl/commit/a21bf3d33838 11:48:12 03Aidan Holm02 07[directn-widget] * 0.26-a0-375-g353da18: Remove redundant redraws in direction chooser 10(4 hours ago, 1 file, 2+ 13-) 13https://github.com/crawl/crawl/commit/353da187aba9 11:52:23 04Build failed for 08directn-widget @ 353da187 06https://github.com/crawl/crawl/actions/runs/174724002 11:57:51 03Aidan Holm02 07[directn-widget] * 0.26-a0-367-g00c3bee: Fix widgets having a size of 0x0 by default 10(10 hours ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/00c3beed2264 11:57:51 03Aidan Holm02 07[directn-widget] * 0.26-a0-368-gff71904: Extract ui::MouseEvent -> wm_mouse_event conversion 10(29 minutes ago, 3 files, 20+ 13-) 13https://github.com/crawl/crawl/commit/ff7190430c3a 11:57:51 03Aidan Holm02 07[directn-widget] * 0.26-a0-369-ga4a24b8: Reimplement direction chooser UI as a widget 10(7 hours ago, 6 files, 143+ 67-) 13https://github.com/crawl/crawl/commit/a4a24b885d62 11:57:51 03Aidan Holm02 07[directn-widget] * 0.26-a0-370-ge4691e7: Don't force screen update when printing stats 10(5 hours ago, 10 files, 14+ 11-) 13https://github.com/crawl/crawl/commit/e4691e74325d 11:57:51 03Aidan Holm02 07[directn-widget] * 0.26-a0-371-ga64ec3d: Don't force screen update in viewwindow 10(5 hours ago, 39 files, 90+ 3-) 13https://github.com/crawl/crawl/commit/a64ec3d783e0 11:57:51 03Aidan Holm02 07[directn-widget] * 0.26-a0-372-ge3431b7: Remove redraw_console_sidebar() hack 10(4 hours ago, 3 files, 0+ 53-) 13https://github.com/crawl/crawl/commit/e3431b71010d 11:57:51 03Aidan Holm02 07[directn-widget] * 0.26-a0-373-gcd99d21: Don't force screen update in redraw_screen 10(4 hours ago, 49 files, 171+ 10-) 13https://github.com/crawl/crawl/commit/cd99d216ee8e 11:57:51 03Aidan Holm02 07[directn-widget] * 0.26-a0-374-gd56f408: Don't update screen when exiting last UI layout 10(4 hours ago, 1 file, 0+ 3-) 13https://github.com/crawl/crawl/commit/d56f408fb4c7 11:57:51 03Aidan Holm02 07[directn-widget] * 0.26-a0-375-gcfc05e4: Remove redundant redraws in direction chooser 10(5 hours ago, 1 file, 2+ 13-) 13https://github.com/crawl/crawl/commit/cfc05e450520 12:02:10 hm, i don't suppose there's a mailing list or anything similar for forks 12:02:25 04Build failed for 08directn-widget @ cfc05e45 06https://github.com/crawl/crawl/actions/runs/174728178 12:02:58 it occurs to me that those commits splitting update_screen() out of viewwindow() and redraw_screen() will cause some very bizarre bugs if merged without manual checks 12:05:17 for all forks? for specific forks? 12:05:42 i guess neither exists, probably 12:08:27 well, any that have added calls to redraw_screen or viewwindow, since they will no longer update the screen 12:13:46 huh, i never realized direction_chooser on webtiles doesn't support mouse control at all 12:13:54 oh well, one less thing to port 12:14:24 03advil02 07* 0.26-a0-367-gaa1b24a: Don't log initfile errors on webtiles (+refactoring) 10(11 minutes ago, 15 files, 254+ 197-) 13https://github.com/crawl/crawl/commit/aa1b24a277f9 12:14:49 The build has errored. (directn-widget - 353da18 #14007 : Aidan Holm): https://travis-ci.org/crawl/crawl/builds/709732231 12:14:58 03Aidan Holm02 07[directn-widget] * 0.26-a0-376-gbec29db: fixup! Reimplement direction chooser UI as a widget 10(26 seconds ago, 1 file, 2+ 0-) 13https://github.com/crawl/crawl/commit/bec29dbb432b 12:14:58 03Aidan Holm02 07[directn-widget] * 0.26-a0-377-gfdc2635: fixup! Extract ui::MouseEvent -> wm_mouse_event conversion 10(17 seconds ago, 2 files, 4+ 0-) 13https://github.com/crawl/crawl/commit/fdc2635644c0 12:15:25 there's something I think hellmonk once said, which is (paraphrasing) that if you want to maintain a fork you still have to pay the crawlcode cost 12:15:25 and I kinda agree with this 12:15:55 but if you wanted to talk to a set of fork authors at once there is the #dcss-variants channel on the roguelikes discord 12:17:32 reminded to finally do aa1b24a by the mention of hda earlier 12:17:58 the crawlcode cost is probably significantly higher for forks 12:18:19 hm, setting up a discord seems like a bit too much effort for me right now 12:24:53 setting up a discord? You mean just a discord account? 12:25:07 pretty easy to register one and use periodically 12:25:43 The build was broken. (directn-widget - cfc05e4 #14008 : Aidan Holm): https://travis-ci.org/crawl/crawl/builds/709733830 12:25:52 I think all prominent fork authors do check said #dcss-variants channgel on roguelikes discord if you just want a line of communication 12:25:55 yeah, you can do it as just a webapp 12:26:28 Unstable branch on crawl.kelbi.org updated to: 0.26-a0-367-gaa1b24a277 (34) 12:26:28 alternately you could make something in the crawl repo; a wiki page for fork issues? 12:28:12 advil: Do you have any knowledge on whether running tornado 6 with SSL is reasonable/stable, unlike how it was for the old 2.x tornado we used? 12:28:56 I know people were using nginx with reverse proxy, which I could certainly set up; haven't set up the web server side of things yet, and was planning the reverse proxy approach, but I realized that might be more necessary for old tornado versions 12:29:09 i do mean just an account, but I'm extremely account-weary at the moment 12:29:26 gammafunk: alexjurkiewicz is running 6 with ssl, it's worked for him 12:29:32 oh, ok 12:29:37 thought he had been using a reverse proxy as well 12:30:01 probably, it seems like something alexjurkiewicz would do 12:30:01 aidanh: understandable; I think fork authors would appreciate some kind of resource for heads-up about changes, even if it wasn't a line of chat 12:30:28 alexjurkiewicz: are you running ssl directly through webtiles these days, or using a reverse proxy? 12:30:44 also, I will say that I have a theory that a big portion of why people find cko so fast (vs other heavy servers in similar geographical locations) is it has a really good proxy setup 12:31:03 could be, although I heard a lot of complaints about cko recently 12:31:13 might have just been tournament overload somehow though 12:32:00 and these complaints were not exactly well-documented, just random grumblings in discord 12:32:00 I know it got resized too low at one point after t 12:32:03 maybe that's what it was 12:33:31 but I know that even people in europe sometimes find it has really good latency 12:34:54 i'm not sure how the proxy would help with that 12:35:23 it's not hitting tornado for bulk downloads? 12:35:30 at least that's how I'm pretty sure it's set up 12:35:59 most of the time there would be no static files being served i imagine 12:36:16 from looking at logs, I'm not sure how true that is 12:36:16 well i guess it depends on the view rate for the main page 12:36:46 it's definitely true that serving static files from tornado is not optimal 12:38:50 my impression is that tornado re-serves all sorts of things a lot 12:39:39 okay, not good yeah 12:40:00 I haven't really tried to quantify this though 12:41:57 here's a grep for "Socket closed" on CAO for ~ the last 4 days, which gives total send data: https://pastebin.com/ZUgAQtDm 12:42:23 tbh not really sure whether that's bad or not for current web standards 12:43:03 I also don't offhand know if the size number is with or without compression 12:45:37 actually, I don't even know if that includes static data 12:46:49 that looks like the total sent across each websocket, which wouldn't include static data 12:47:53 oh, I htink we actually have GET logging suppressed by default because it's so spammy 12:50:17 heh 12:51:09 alexjurkiewicz random observation, I think your weekly challenge might be broken right now because it is --sk (obv is slated to change in a few hours anyways, so may not be worth doing anything for) 12:52:03 oh, in other --sk news, and equally random, cwz has trunk updated to have palentonga but not so recent as to have charms removal 12:52:07 !hw * pask 12:52:17 !hs * pask 12:52:17 No keyword 'pask' 12:52:22 !hs * char=pask 12:52:23 150. Yermak the Slayer (L27 PaSk of Cheibriados), escaped with the Orb and 15 runes on 2020-07-19 11:09:35, with 58310663 points after 24700 turns and 13:05:39. 12:52:24 so yermak did that on cwz 12:52:39 which was that game that got griefed by the chat message troll repeatedly 12:57:22 it seems both cao and cko have cache busting via a query parameter 12:57:38 not sure what the relative traffic rates or cpu utilization rates are like either 13:01:07 oh, one major difference is that cko sets cache-control: public, which allows intermediate proxies to cache static resources 13:01:51 cao is actually missing a bunch of headers like the content type; i'm surprised it works at all 13:09:01 the http headers sent by WebTiles? I thought cao was upgraded to the new webtiles codebase 13:12:29 cao is on tornado 4.5.3 13:12:54 which is still fairly old; advil would know more about the reason for that (probably python 3?) 13:13:23 oh, right 13:13:48 it can't run tornado 6 because of no py3, it can't run tornado 5 because the py2 version is too old 13:14:28 (although, if you're running under py2, I'm pretty sure tornado 5 is actually slower than 4, it's kind of an interim version) 13:16:53 although I did get py3 working in a chroot! (but haven't yet been brave enough to try to use that for the server...) 13:20:52 Unstable branch on crawl.akrasiac.org updated to: 0.26-a0-367-gaa1b24a (34) 13:43:50 The build passed. (directn-widget - fdc2635 #14010 : Aidan Holm): https://travis-ci.org/crawl/crawl/builds/709737051 14:01:43 03Aidan Holm02 07* 0.26-a0-368-g9df63ee: Fix widgets having a size of 0x0 by default 10(12 hours ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/9df63ee947b5 14:01:43 03Aidan Holm02 07* 0.26-a0-369-g8d37012: Extract ui::MouseEvent -> wm_mouse_event conversion 10(3 hours ago, 3 files, 24+ 13-) 13https://github.com/crawl/crawl/commit/8d3701215b01 14:01:43 03Aidan Holm02 07* 0.26-a0-370-g38af612: Reimplement direction chooser UI as a widget 10(10 hours ago, 6 files, 145+ 67-) 13https://github.com/crawl/crawl/commit/38af612a8a6b 14:01:43 03Aidan Holm02 07* 0.26-a0-371-g3ebe85e: Don't force screen update in print_stats 10(7 hours ago, 10 files, 14+ 11-) 13https://github.com/crawl/crawl/commit/3ebe85eda627 14:01:43 03Aidan Holm02 07* 0.26-a0-372-gc6476c8: Don't force screen update in viewwindow 10(7 hours ago, 39 files, 90+ 3-) 13https://github.com/crawl/crawl/commit/c6476c8bcdda 14:01:43 03Aidan Holm02 07* 0.26-a0-373-g72aac4d: Remove redraw_console_sidebar() hack 10(6 hours ago, 3 files, 0+ 53-) 13https://github.com/crawl/crawl/commit/72aac4d990ab 14:01:43 03Aidan Holm02 07* 0.26-a0-374-g5a0d02c: Don't force screen update in redraw_screen 10(6 hours ago, 49 files, 171+ 10-) 13https://github.com/crawl/crawl/commit/5a0d02cfaed7 14:01:43 03Aidan Holm02 07* 0.26-a0-375-gf9beda2: Don't update screen when exiting last UI layout 10(6 hours ago, 1 file, 0+ 3-) 13https://github.com/crawl/crawl/commit/f9beda28d27b 14:01:43 03Aidan Holm02 07* 0.26-a0-376-g13fbb91: Remove redundant redraws in direction chooser 10(7 hours ago, 1 file, 2+ 13-) 13https://github.com/crawl/crawl/commit/13fbb91ecec6 14:04:29 alexjurkiewicz, Pinkbeast: those commits that split out update_screen() will require manually searching for any viewwindow/redraw_screen/print_stats calls that your forks have added 14:06:47 Thanks (I'll make a note since at the current pace we'll hit 0.26 in about 2025...) 14:09:05 heh 14:27:04 Unstable branch on crawl.kelbi.org updated to: 0.26-a0-376-g13fbb91ece (34) 15:09:13 btw in case anyone hadn't noticed, it appears that crawl (and sequell) made the cut for the arctic vault 15:11:37 well, not that it looks like it was that selective 15:12:55 but we all get a badge! 15:22:35 amazing 18:23:12 Unstable branch on underhound.eu updated to: 0.26-a0-376-g13fbb91ece (34) 20:36:31 just to check 20:36:35 does ghoul still have the rot gimmick 20:36:45 or does it currently have no clock like every other species 20:47:04 advil: funnily enough the rng picked sk for last week and delver for this week 20:47:33 gammafunk: using a reverse proxy, because I want to serve things other than webtiles on the same domain name (eg morgues, etc) 20:48:27 i guess this means it's time for bloatcrawl 3 to start development 21:06:50 03PleasingFungus02 07* 0.26-a0-377-g2951c22: Never anger allies 10(8 hours ago, 12 files, 64+ 229-) 13https://github.com/crawl/crawl/commit/2951c229262b 21:11:13 big 21:12:46 related: wand of clouds is marked as forbiden under zin, but only because one of the potential cloud types causes penance. However there are secret breakpoints below which this cloud type will never generate. Can Zin suppress this cloud type silently 21:12:46 after writing that out maybe i'll just PR it myself 21:15:29 we've been moving away from having gods silently make spells/items work differently like that, i'd rather try and stay consistent in that direction personally 21:17:23 i guess clouds having a breakpoint for negative energy makes it more awkward there but maybe it'd be better to remove the breakpoint and still have a low chance of the strong cloud types at low power 21:18:17 TSO prevents poison sting mutation. I guess lore-wise you can flavour that as "self control"? 21:22:25 is there some way to calculate the odds of random picker results at certain input powers? 21:30:06 Unstable branch on crawl.kelbi.org updated to: 0.26-a0-377-g2951c22926 (34) 23:09:46 ^status 23:09:46 29 Crawlers. CBRO disk usage=93% (135GB) | RAM usage=50% (4GB)| uptime/CPU= 23:09:45 up 54 days, 17:45, 2 users, load average: 0.42, 0.47, 0.43 (4 Cores) http://status.berotato.org 23:20:26 Why are sacrifices GHIJ ? like what's the reasoning behind that position? 23:26:58 03PleasingFungus02 07* 0.26-a0-378-g7e20f67: Don't make allies hostile when joining a god 10(73 seconds ago, 3 files, 1+ 38-) 13https://github.com/crawl/crawl/commit/7e20f673c034 23:34:27 woohoo, TSO demons finally? 23:34:34 the summoning god is here! 23:38:34 no TSO still puts you under pennace if you summon demons 23:38:44 the only difference is the demons will (probably) be friendly 23:44:39 ah, rats 23:47:50 Unstable branch on crawl.kelbi.org updated to: 0.26-a0-378-g7e20f673c0 (34) 23:55:48 Rats will be fine to summon under TSO still. 23:59:34 hell rats won't 23:59:52 yeah! I win pedantic correction