00:17:10 -!- Amnesiac_ is now known as Amnesiac 00:28:31 grumblor (L13 SpGl) ASSERT(mgrd(mons->pos()) == s) in 'show.cc' at line 275 failed. (Lair:3) 01:55:31 -!- Psymania_ is now known as Psymania 03:41:43 -!- EuphOria is now known as OICU812 03:44:40 03Aidan Holm02 07[py2019] * 0.25-a0-592-g443c2f7: Remove remember-me checkbox 10(7 days ago, 2 files, 1+ 23-) 13https://github.com/crawl/crawl/commit/443c2f7d93e5 03:44:40 03advil02 {Aidan Holm} 07[py2019] * 0.25-a0-593-g31b98c6: Preliminary support for Tornado 5+ 10(6 months ago, 10 files, 97+ 97-) 13https://github.com/crawl/crawl/commit/31b98c656dbc 03:44:40 03advil02 {Aidan Holm} 07[py2019] * 0.25-a0-594-g86d8afd: Updates for python 3 compatibility 10(6 months ago, 6 files, 74+ 64-) 13https://github.com/crawl/crawl/commit/86d8afde4904 03:44:40 03advil02 {Aidan Holm} 07[py2019] * 0.25-a0-595-g5a12d5d: Webtiles: minimal type annotations to pass static type checking 10(6 months ago, 6 files, 27+ 6-) 13https://github.com/crawl/crawl/commit/5a12d5dcd8c2 03:44:40 03advil02 {Aidan Holm} 07[py2019] * 0.25-a0-596-g1885c73: Webtiles: static type annotations for low-level messaging 10(6 months ago, 3 files, 55+ 42-) 13https://github.com/crawl/crawl/commit/1885c7323028 03:44:40 03advil02 {Aidan Holm} 07[py2019] * 0.25-a0-597-g88d3183: A few small tweaks + one more byte/str fix 10(3 weeks ago, 3 files, 12+ 4-) 13https://github.com/crawl/crawl/commit/88d31831b636 03:45:35 advil: I've rebased py2019 and fixed a missing IOLoop import 03:45:41 03Aidan Holm02 07* 0.25-a0-592-g443c2f7: Remove remember-me checkbox 10(7 days ago, 2 files, 1+ 23-) 13https://github.com/crawl/crawl/commit/443c2f7d93e5 03:48:46 Monster database of master branch on crawl.develz.org updated to: 0.24-a0-443-g80245de385 04:18:55 03advil02 {Aidan Holm} 07[webserver-dev-env] * 0.25-a0-593-g31b98c6: Preliminary support for Tornado 5+ 10(6 months ago, 10 files, 97+ 97-) 13https://github.com/crawl/crawl/commit/31b98c656dbc 04:18:55 03advil02 {Aidan Holm} 07[webserver-dev-env] * 0.25-a0-594-g86d8afd: Updates for python 3 compatibility 10(6 months ago, 6 files, 74+ 64-) 13https://github.com/crawl/crawl/commit/86d8afde4904 04:18:55 03advil02 {Aidan Holm} 07[webserver-dev-env] * 0.25-a0-595-g5a12d5d: Webtiles: minimal type annotations to pass static type checking 10(6 months ago, 6 files, 27+ 6-) 13https://github.com/crawl/crawl/commit/5a12d5dcd8c2 04:18:55 03advil02 {Aidan Holm} 07[webserver-dev-env] * 0.25-a0-596-g1885c73: Webtiles: static type annotations for low-level messaging 10(6 months ago, 3 files, 55+ 42-) 13https://github.com/crawl/crawl/commit/1885c7323028 04:18:55 03advil02 {Aidan Holm} 07[webserver-dev-env] * 0.25-a0-597-g88d3183: A few small tweaks + one more byte/str fix 10(3 weeks ago, 3 files, 12+ 4-) 13https://github.com/crawl/crawl/commit/88d31831b636 04:18:55 03Aidan Holm02 07[webserver-dev-env] * 0.25-a0-598-g2a08999: Add requirements.txt tools, formatting, and linting 10(13 days ago, 9 files, 172+ 0-) 13https://github.com/crawl/crawl/commit/2a089998a254 04:18:55 03Aidan Holm02 07[webserver-dev-env] * 0.25-a0-599-gc7cac34: Fix lint errors 10(7 days ago, 1 file, 104+ 52-) 13https://github.com/crawl/crawl/commit/c7cac34b2eb8 04:18:55 03Aidan Holm02 07[webserver-dev-env] * 0.25-a0-600-g2b597b6: Refactor userdb.py with a context manager 10(7 days ago, 1 file, 136+ 213-) 13https://github.com/crawl/crawl/commit/2b597b670ee5 04:18:55 03Aidan Holm02 07[webserver-dev-env] * 0.25-a0-601-g4283e85: Added User class 10(7 days ago, 2 files, 24+ 4-) 13https://github.com/crawl/crawl/commit/4283e851184c 04:18:55 03Aidan Holm02 07[webserver-dev-env] * 0.25-a0-602-gfed15be: Add base request handler classes 10(4 days ago, 4 files, 21+ 4-) 13https://github.com/crawl/crawl/commit/fed15be3a235 04:18:55 ... and 8 more commits 04:18:58 Branch pull/1306 updated to be equal with webserver-dev-env: 13https://github.com/crawl/crawl/pull/1306 05:20:36 The build was broken. (webserver-dev-env - e7cb76c #12826 : Aidan Holm): https://travis-ci.org/crawl/crawl/builds/662326812 06:02:23 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-592-g443c2f7d93 (34) 07:01:12 advil: is there anything stopping py2019 from being merged btw? it only changes py files, so it should be relatively safe 11:17:44 aidanh there are a bunch of potential issues, and I wanted to be available to monitor it 11:17:59 may have time this week, what with spring break + coronavirus 11:18:45 in principle it is all designed to be backwards compatible though 11:19:29 yeah, it's quite fortunate that tornado provides that level of backwards compatibility 11:19:57 fwiw it's working fine on ubuntu 12.04 11:20:34 which is python 2.7.3 11:20:34 yeah I tested it very carefully locally 11:20:34 the real question as always is what happens when I flip the switch on CAO 11:21:25 i'm guessing cao has gone a while without a webserver restart 11:21:26 I did figure out the biggest issue there which is that CAO is *constantly* tripping the blocking timeout, which newer tornados don't support 11:21:29 haha yes 11:21:42 though not really that long 11:21:54 I did restart CAO webtiles maybe in Nov or Dec? 11:22:04 oh that's not too bad 11:22:06 needed to happen because of some of the ui changes 11:22:16 when I did that it had been like years though or something 11:23:08 weekly automated restarts might not be a bad idea 11:23:33 one thing we need is something that announces an upcoming restart to active players 11:24:26 hm, true 11:25:05 i've rebased my webserver-dev-env branch on top of py2019, btw 11:25:09 yeah I just saw 11:25:15 actually I think I'm going to just merge it today 11:25:24 \o/ 11:25:26 it'll give me something to focus on that is not the world burning 11:25:58 no kidding 11:25:58 are you on wfh yet? 11:26:19 I have to learn how to teach online in the next week 11:26:46 optional wfh here, since the situation is pretty mild so far 11:26:59 mild is good 11:27:29 we get our hands sprayed going into the office, and into shopping centres, restaurants, etc 11:27:35 it seems mild here but the fact is no one knows because there's effectively no testing here 11:28:01 right, the incubation period is pretty frightening 11:28:57 yeah 11:29:05 and yeah, I've heard about the lack of testing and other general failures of coordination over there; it's quite concerning 11:30:13 I talked to a friend of mine in italy earlier this week and it was ... not encouraging 11:34:03 03advil02 07* 0.25-a0-593-g6dcb09c: Preliminary support for Tornado 5+ 10(6 months ago, 10 files, 96+ 97-) 13https://github.com/crawl/crawl/commit/6dcb09c756f2 11:34:03 03advil02 07* 0.25-a0-594-g58c0c3d: Updates for python 3 compatibility 10(6 months ago, 6 files, 74+ 64-) 13https://github.com/crawl/crawl/commit/58c0c3d08e25 11:34:03 03advil02 07* 0.25-a0-595-g9a49a99: Webtiles: minimal type annotations to pass static type checking 10(6 months ago, 6 files, 27+ 6-) 13https://github.com/crawl/crawl/commit/9a49a99f9af0 11:34:03 03advil02 07* 0.25-a0-596-gf09cd5c: Webtiles: static type annotations for low-level messaging 10(6 months ago, 3 files, 55+ 42-) 13https://github.com/crawl/crawl/commit/f09cd5ce1912 11:34:03 03advil02 07* 0.25-a0-597-gb30770a: A few small tweaks + one more byte/str fix 10(3 weeks ago, 3 files, 12+ 4-) 13https://github.com/crawl/crawl/commit/b30770a0c043 11:35:29 so as long as everyone sticks to python 2.7 with the same version of tornado, behavior *should* be identical 11:36:14 I don't advise trying to upgrade highest traffic servers (on either dimension) yet, and especially not to tornado 6, though it should in principle work 11:36:28 awesome 11:37:16 also should not have an effect without manually restarting webtiles 11:37:23 I will aim to restart cao sometime today, and if that goes badly I guess I'll have to figure out a more cautious approach 11:37:52 i've started adding some test cases, so it should be possible to port to py3 / tornado 6 relatively worry-free 11:38:15 well, I expect I haven 11:38:20 t found every byte/str issue 11:38:53 one thing I noticed with that very last commit is that the static type stuff I added won't catch type problems with callbacks 11:39:07 and for CAO the blocking issue is a real problem 11:39:20 the blocking issue being what, sorry? 11:39:45 older versions of tornado support setting a timeout that breaks out of blocking calls with an error 11:39:46 tornado 6 does not support this 11:39:57 the timeout is by default at 500ms, and it trips constantly on CAO 11:40:12 right, I remember seeing some of that 11:40:23 as far as I can tell it is because of extremely slow disk access (though I've never found an instance where e.g. a write fails) 11:41:33 that's concerning; is cao usually fairly laggy for players? 11:41:49 based on the time it can take to do simple things like `ls` in the wrong directory, I think if certain disk-related calls were not timed out it might block for many seconds, routinely 11:41:56 yes, it is 11:42:01 er, it's concerning 11:42:07 I think it's relatively laggy but not unplayable 11:42:19 I don't think these disk calls happen much while playing normally 11:42:40 any way to collect tracebacks whenever this happens? 11:42:57 one case seems to be random access to something that is not currently in cache 11:42:59 I've tried 11:43:19 but nearly everything I've come up with for debugging also creates a fair amount of lag 11:43:39 (it definitely makes me worry about the lifespan of the disk in question...) 11:44:14 though I think it might just be about the sheer number of files on the disk 11:46:00 bugger 11:46:32 even for hard disks, multiple seconds sounds excessive 11:47:30 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-597-gb30770a0c0 (34) 11:47:47 here's a sample: https://pastebin.com/yKG3MwTM 11:48:12 at some point I did actually check what it's calling there, can't remember now 11:48:21 whatever it was actually does seem to happen 11:51:26 2020-03-14 11:50:13,787 WARN: #0 Didn't know how to handle msg: ui_state_sync 11:51:33 yeah that's harmless 11:51:38 which I was getting before the commit as well, so I don't know if … ok 11:52:00 I tracked down what's happening there, it's just a message that gets sent from js side after the socket closes 11:52:06 I also had one run die with an error; haven't reproduced with the updates yet, have tested with bionic's python 3 11:52:07 I didn't figure out a fix at the time 11:52:29 this si my first time running webtiles on this particular machine, and I may drop it because linking takes noticeably longer 11:52:48 looks like it's lagging while copying the player's RC files? 11:52:56 yeah, that sounds right 11:53:10 oh one hilarious (annoying) side effect is that it routinely times out while rotating logs 11:53:15 raise RuntimeError("IOLoop is already running") 11:53:33 File "webserver/server.py", line 259, in 11:53:50 geekosaur do you mean to say that you get that exception when running with current trunk? 11:53:51 (can't do multiline pastes in this client so cutting down to the minimum) 11:54:02 in fact it looks like it seeks to the rc directory, checks for the folder, then seeks back to the default rc directory to copy it 11:54:02 that was actually before your updates 11:54:11 so perhaps just cp --no-clobber would be faster 11:55:03 ah this is sounding familiar 11:55:18 I'll keep playing webtiles and see if I can reproduce it with the current code 11:57:07 that crash happens on server shutdown fwiw; I had already exited the game (well, died to an orc priest, sigh) and it should have had plenty of time to exit 11:59:53 oh, yeah the shutdown stuff was pretty buggy 12:00:11 it should actually be fixed as of the py2019 branch, but let me know if it's not 12:02:14 also I get a spurious "Login failed" on first loading the lobby, with local webtiles. which is probably just from hitting reload on /#lobby instead of starting from / 12:03:14 the auth handling in general is pretty wonky 12:13:58 Unstable branch on crawl.akrasiac.org updated to: 0.25-a0-597-gb30770a (34) 12:23:28 Stable (0.23) branch on underhound.eu updated to: 0.23.1-93-ge536e68a2c 12:47:43 does anyone talk to the cxc admin or know how to contact? 12:48:03 advil: They used to be on Discord, checking... 12:48:06 should invite them to the serveradmins team 12:48:10 I'm not on the cxc discord 12:48:23 I mean, on the big Discord - I'm not either 12:48:37 "Nanamix", but they show as not logged in right now. 12:54:29 I guess someone should also see if the LLD / webzook admins want to join the github team? 12:55:23 maybe gammafunk knows how to contact them 13:04:23 I think maybe I should restart CAO late at night, it's really busy right now 13:10:06 aidanh would you be able to update the draft security advisory with the further changes you made? I can do it but it'll take me a bit longer to dig them up 13:10:49 afaik we are considering that vulnerability patched after those, right? 13:11:40 at least, patched in terms of code 13:12:07 ... to be clear, is this _another_ one, or the one from a couple of weeks ago? 13:12:50 the one from a couple weeks ago 13:12:56 we never published the advisory 13:14:50 sure thing 14:05:15 Blink cast by Erolcha The Cunning didn't narrow her set of spellbooks. 13https://crawl.develz.org/mantis/view.php?id=12222 by Yermak 14:42:39 -!- Tiobot is now known as Guest22961 14:43:56 -!- broquain1 is now known as broquaint 16:31:36 -!- Psymania_ is now known as Psymania 16:52:50 Spell name in ghost description window is cut-off 13https://crawl.develz.org/mantis/view.php?id=12223 by Goratrix 18:41:18 in case anyone asks about this: we have disabled telnet on CAO, probably forever 18:42:23 but if there are players still using it that this inconveniences somehow, I don't mind hearing from them (I checked with r.ax before doing it but I'm the one to blame) 18:50:54 I guess CDO is the only server still running telnet, but on a non-standard port 19:22:53 Unstable branch on underhound.eu updated to: 0.25-a0-597-gb30770a0c0 (34) 19:53:23 oh, didn't realize CDO did run telnet 19:53:34 I assume nap.kin has a good reason for that 20:02:10 New branch created: pull/1316 (1 commit) 13https://github.com/crawl/crawl/pull/1316 20:02:10 03Skrybe02 {GitHub} 07https://github.com/crawl/crawl/pull/1316 * 0.25-a0-598-gbd11b89: Update layout_cellular.des 10(14 minutes ago, 1 file, 6+ 5-) 13https://github.com/crawl/crawl/commit/bd11b897b217 23:10:55 advil: is cao running on that patched version of tornado the docs refer to? 23:12:50 i.e. edlothiol's patch for the connection_timeout thing 23:12:58 if so, i've got some bad news; that's tornado 2.4, not 3.2.2, and it doesn't have the IOLoop.current() 23:17:21 03Aidan Holm02 07* 0.25-a0-598-gc9c9e4c: Add missing IOLoop import 10(23 seconds ago, 1 file, 1+ 0-) 13https://github.com/crawl/crawl/commit/c9c9e4cfe1d3 23:20:10 hrm, was that argument later added to tornado for the relevant call? 23:20:33 yeah, it looks like 3.2.2 works fine 23:21:17 just something to be aware of if you get bizarre errors about an IOLoop AttributeError 23:21:58 hrm, I can check the log right now 23:22:36 i don't think advil has restarted cao yet 23:23:33 I see 23:25:13 if there's a PYTHONPATH=/home/crawl-dev/tornado in /etc/init.d/webtiles that'd probably need to be removed as well 23:25:39 as well as running `pip install tornado==3.2.2`, that is 23:27:35 oh, and that init.d script comes from dgamelaunch-config, in utils/webtiles 23:28:02 yep; I'll not be doing this upgrade, anyhow, since advil has a far better idea of what's going on 23:29:37 I was going to ask what's special about tornado 3.2.2, but I'm remembering that we don't have python3 on cao and hence that's probably the latest tornado still compatible with python2? 23:30:19 afaik nothing, other than it's known to work with the current webtiles code 23:30:34 no, tornado 6 should be comptible with py2 23:31:00 oh ok 23:32:43 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-598-gc9c9e4cfe1 (34)