00:02:19 Rest key doesn't work at all when you're sticky flamed. Wish: make it spend one turn. 01:01:15 -!- amalloy_ is now known as amalloy 01:31:12 -!- signop1 is now known as signop 01:38:18 Unstable branch on crawl.develz.org updated to: 0.26-a0-154-g3d63440f56 (34) 02:55:31 Hello 02:56:23 My trunk game on CXC keeps crashing - is there anyone who could take a look? 02:56:23 I get a message about the websocket closing suddenly 03:15:23 ??cxc 03:15:24 cXc[1/2]: Server in France: CRAWL.XTAHUA.COM -- ssh port 22, username crawl, uses CAO key (available at http://CRAWL.XTAHUA.COM/crawl/keys/). Webtiles address: http://CRAWL.XTAHUA.COM Discord server https://discord.gg/SzXGZNS 03:15:36 try contacting the cxc admins on their discord tilp_ 03:30:33 Fork (bcrawl) on crawl.kelbi.org updated to: 0.23-a0-3355-g331662a87b 06:40:01 -!- amalloy is now known as amalloy_ 07:58:35 -!- dolemite8 is now known as dolemite7 11:26:26 advil: lol at that "why does CMD_EVOKE call this function" comment 11:26:43 as I recall you were the one who hacked that together 11:28:15 using a raii stack for cursors is a good idea 11:28:56 it's a shame the main game rendering is so incredibly tangled together: refactoring the rest of the ui was a cakewalk in comparison 11:28:57 it looks like I moved that call from elsewhere 11:29:04 it's a good idea but that branch is currently a disaster 11:29:09 the cursor position is basically never valid 11:29:38 so only slightly worse than regular crawl; ship it! 11:29:52 !source redraw_console_sidebar 11:29:53 1/1. https://github.com/crawl/crawl/blob/master/crawl-ref/source/output.cc#L1483 11:30:29 %git bb52cfba1566 11:30:29 07advil02 * 0.23-a0-52-gbb52cfb: Don't do extra map redraws in direction chooser 10(1 year, 11 months ago, 3 files, 39+ 1-) 13https://github.com/crawl/crawl/commit/bb52cfba1566 11:30:36 guess that's the reason 11:31:28 though in the original commit it was used fairly generally 11:31:35 %git d13e5294dc 11:31:35 07advil02 * 0.23-a0-57-gd13e529: Limit a redraw to the special case it is fixing (gammafunk) 10(1 year, 11 months ago, 2 files, 6+ 5-) 13https://github.com/crawl/crawl/commit/d13e5294dc3e 11:31:45 guess I can delete the comment ;-) 11:38:10 I thought ncurses was supposed to limit screen writes via diffing 11:39:16 not sure why that isn't working in that case 11:39:32 you mean why was the original version flickering? 11:41:20 I would guess that there was a non-trivial diff because of an undiscovered bug, but I can't remember the details of what the "display problems" were any more 11:42:54 %git 2f468e5a 11:42:54 07Aidan Holm02 * 0.22-a0-795-g2f468e5: Work around direction-chooser console rendering bug 10(1 year, 11 months ago, 1 file, 4+ 0-) 13https://github.com/crawl/crawl/commit/2f468e5a82e4 11:43:55 the original desc does make it seem like there's a bug that neither of us ever directly fixed if it's even possible for an item description showing up in the stats region 11:44:46 possibly not irrelevant to what I'm trying to do in this branch, since some crashes that I think I've actually fixed result from the cursor position spuriously being in the stats region 11:53:37 Webtiles server stopped. 12:08:04 it's less a bug, more that the direction_chooser stuff only ever expects to be called when the main view is shown, so it never bothers to clear the sidebar 12:08:22 *redraw the sidebar 12:21:28 Webtiles server started. 12:28:48 digitalocean is offering more instance sizes at the low price point, i can pay $15 now instead of $20 for 2GB of memory instead of 4 12:29:10 i'm thinking that should be OK for a 40-50 player count, i think the cpu is the bottleneck anyway 12:36:04 should be fine I reckon 12:46:15 floraline: how many cores does cko have? 12:47:14 or vcpus, if it's a virtual environment (not sure if other virtual servers use something analogous to what amazon's ec2 uses) 12:52:21 huh I would have guessed that something like crawl would not work on a shared cpu droplet because the cpu is too continuous? 12:54:01 gammafunk based on what floraline just said I would guess cko is on a 2vCPU droplet 12:54:17 under "standard droplets" https://www.digitalocean.com/pricing/ 12:55:22 surprised you can host so many concurrent users with that much cpu 13:02:01 yes 2 v cpus 13:03:28 i resize it to 4 during the tournament 13:05:51 i did have a suspicion that crawl might intrude on other guests too much, but DO hasn't said anything about it so it must not be a problem 14:19:16 New branch created: tile-ray-refactor (1 commit) 13https://github.com/crawl/crawl/tree/tile-ray-refactor 14:19:16 03Aidan Holm02 07[tile-ray-refactor] * 0.26-a0-155-gae42f58: Simplify tile_draw_rays() 10(3 minutes ago, 3 files, 11+ 13-) 13https://github.com/crawl/crawl/commit/ae42f58cee8d 14:22:34 aidanh: someone mentioned offhand that ignition no longer has animations and that monsters just die; haven't looked into it, but does that sound like something that could have broken? 14:23:02 this wasn't someone who had just turned off animations in their RC or anything like that; was after the cloud visibility fix 14:23:05 yeah, probably 14:23:32 i refactored tile overlays recently 14:23:51 which is what things like infestation, ignition, etc use 14:24:00 i'll look into it tomorrow 14:24:03 thx 14:24:24 np, thanks for letting me know 14:48:42 pretty sure the console animations were working? 14:48:58 (though ofc they are one of the things that currently crashes in my branch) 15:07:02 The build passed. (tile-ray-refactor - ae42f58 #13706 : Aidan Holm): https://travis-ci.org/crawl/crawl/builds/704033975 17:30:43 -!- toblu1 is now known as toblu 18:20:24 crawl is pretty light on CPU, fwiw 18:20:40 it's not really continuous cpu usage because per-player input is so spiky, eg most of the time is waiting for input 18:21:02 even with 50 users you see per-second minimums and maximums of 0/100% 18:28:42 interesting 18:29:31 I guess I only see cao, and I mostly look at its cpu when there's a problem or I'm trying to do something else expensive on the server 18:34:10 well, I now understand many more things about cursor position; current belief: many of the invalid position issues that have seemed kind of inexplicable are because when you write a symbol in the right-most column of the screen in curses, it moves the cursor to the next row for you, leading to what our internal checks think is an invalid cursor pos 18:34:36 the reason it's always the monster list pane in these errors is because that pane consistently uses the last column, the rest of the sidebar usually doesn't 18:42:39 in retrospect it's not shocking that wrapping is what happens in this case, because where else does the cursor go, but it's a bit of a headache relative to this famous "screen write out of bounds" ASSERT (and all my more involved versions of that) 20:37:03 Branch master updated to be equal with impish: 13https://github.com/crawl/crawl/tree/master 20:37:14 Branch pull/1463 updated to be equal with impish: 13https://github.com/crawl/crawl/pull/1463 20:42:18 The build has errored. (impish - 60f854e #13707 : Nicholas Feinberg): https://travis-ci.org/crawl/crawl/builds/704145569 20:52:42 Hi everyone, I have a question: 20:52:46 How can I register mantis to write a post to tavern? 20:53:25 viktor67: Presently registration is broken, with no fix in sight. 20:54:32 Oh, I see. :( 20:55:54 Branch master updated to be equal with hedgewiz: 13https://github.com/crawl/crawl/tree/master 20:56:05 Branch pull/1466 updated to be equal with hedgewiz: 13https://github.com/crawl/crawl/pull/1466 21:13:37 The build has errored. (hedgewiz - 35b0e8a #13710 : Nicholas Feinberg): https://travis-ci.org/crawl/crawl/builds/704148834 21:40:47 update: resizing to 2GB ram was a mistake 21:40:47 Webtiles server started. 21:41:48 the system immediately ran off the rails after free memory was depleted 21:42:16 i hope no games were affected 21:46:46 Unstable branch on crawl.kelbi.org updated to: 0.26-a0-157-g35b0e8a925 (34) 21:47:18 probably fine 21:48:22 people will have rolled back a bit and that kind of thing, but unless the oom happened literally while a game was midsave actual corruption is unlikely 21:48:59 yeah, I learned when setting up NotSequell that DO droplets are extremely unforgiving wrt out of memory errors 21:49:04 because they do not use swap 21:50:35 -!- Tuxedo[Qyou] is now known as Aux[Qyou] 22:15:45 -!- Aux[Qyou] is now known as Tux[Qyou] 22:30:20 why is my mouse wheel moving me in console 22:32:27 -!- amalloy_ is now known as amalloy 23:19:21 03advil02 07[cursor-region-sanity] * 0.26-a0-156-g281e475: Set cursor region on fakecursorxy 10(14 hours ago, 3 files, 10+ 1-) 13https://github.com/crawl/crawl/commit/281e47575add 23:19:21 03advil02 07[cursor-region-sanity] * 0.26-a0-157-gdb58ac7: Use a smarter cprintf in output.cc 10(75 seconds ago, 5 files, 88+ 11-) 13https://github.com/crawl/crawl/commit/db58ac7b165d 23:33:13 04Build failed for 08cursor-region-sanity @ db58ac7b 06https://github.com/crawl/crawl/actions/runs/154868160 23:46:22 03advil02 07[cursor-region-sanity] * 0.26-a0-157-g0466488: Use a smarter cprintf in output.cc 10(28 minutes ago, 5 files, 97+ 11-) 13https://github.com/crawl/crawl/commit/0466488019ea