00:01:18 doesn't seem to be a traceback generated by the 403 00:02:53 It should be in tornado/web.py 00:03:03 ok, getting somewhere 00:03:07 ooh, is this success 00:03:21 I have tornado 4 working 00:03:48 \o/ 00:04:21 03advil02 07* 0.25-a0-609-g7ee555d: Fix an import 10(35 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/7ee555d2e04e 00:04:39 wow, this Toggle admin panel button 00:04:39 "Please do not use this frivolously!" 00:04:45 I anticipated your desire to tell people about necro spells :-P 00:04:51 indeed 00:05:15 tempting, tempting 00:10:03 %git 734848013 00:10:03 07advil02 * 0.25-a0-601-g7348480: Tornado 2.4 compatibility hacks 10(12 hours ago, 2 files, 35+ 3-) 13https://github.com/crawl/crawl/commit/734848013d98 00:10:51 %git 2bf96b3bd9 00:10:51 07Aidan Holm02 * 0.25-a0-599-g2bf96b3: Fix broken static files base path 10(13 hours ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/2bf96b3bd929 00:14:12 really confused as to why using the old behavior is breaking this 00:17:40 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-609-g7ee555d2e0 (34) 00:18:57 Webtiles server started. 00:19:43 Webtiles server stopped. 00:24:34 doesn't seem to have any bearing on the puzzles being looked at, but have been looking at Tornado docs, they recommend calling static_url() for static content urls, which it seems adds a "v" argument to the querystring which is a hash of the content, I guess it helps with version control 00:26:20 aidanh do you have any idea how this ever worked?? with edlothiel tornado 2.4 I'm getting this exact same problem with "." as the initialization 00:26:27 works fine if I just run it from one directory up 00:29:04 maybe try "./" instead of "." if you're trying to do a path? it's been awhile since I looked at bash conventions, but I think you have problems if you don't end with a "/", it's not to standards.. if you're in a current directory and you want to run something, like "./execute" ? Just a though since it sounds like a directory traversal issue from your last comment 00:31:55 Not having a trailing slash means that its unclear if the path is a file or a directory 00:33:20 Unstable branch on crawl.akrasiac.org updated to: 0.25-a0-609-g7ee555d (34) 00:34:41 03advil02 07* 0.25-a0-610-gd0df705: Fix the ancient tornado compat code 10(46 seconds ago, 2 files, 1+ 1-) 13https://github.com/crawl/crawl/commit/d0df705f03ee 00:34:59 I just ended up overriding the path calculation that really old tornado versions do 00:35:05 the puzzle is that "." was the value it used up until these changes 00:35:39 almost ready to bring it back up I think 00:36:19 maybe then it's due to a different version of python handling stuff differently (?) 00:36:45 the python version isn't what's changed here; the tornado version is 00:36:59 except this is for compatibility with the old tornado version 00:37:03 huh, if there are files in sockets/ while the server is down, I guess I can probably delete them? 00:37:16 esp. if they're all from 3-5 years ago 00:38:10 Unstable branch on crawl.akrasiac.org updated to: 0.25-a0-610-gd0df705 (34) 00:39:29 Webtiles server started. 00:40:24 well, I can start up a game 00:40:24 not going to actualy start one since my winrate... 00:40:36 weird that when spectating someone at char select your mouse cursor selects buttons 00:41:31 game appears to play fine so far (on different account) 00:41:41 yeah I tried out my test account briefly 00:42:07 and spectating seems ok 00:42:40 well, that was mildly more painful than I'd hoped 00:42:44 my game works too - well nice that the new version of Tornado doesn't work with what seems to be a funky parameter (bug) (cough)... why not just have everyone upgrade their tornado version and do away with the compatibility kludge. 00:44:08 and boy do I have some cheese going - flying boots standing in water (crypt3) casting vile clutch on the poor souls at the door and hitting them with sniper.. so cheap :D 00:45:58 thanks for making those upgrades, advil 00:46:09 I guess at some point we really need a dist upgrade on CAO, or something 00:46:12 well, we do want everyone to update their tornado version but I'm trying to avoid a scenario where a server op isn't paying attention and restarts their server to disaster 00:46:52 since these setups are kind of delicate, and the official tornado version is just *so* far behind 00:47:08 (in principle this could be a move from tornado 2.4 to tornado 6, for some servers) 00:47:38 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-610-gd0df705f03 (34) 00:51:12 I really hope after all this that more recent tornado versions actually are faster or the same speed 00:52:19 I saw in the tornado docs in the static file handling section, it actually said to use nginx, but then I guess that would be a bit messy using two web services 00:53:34 yeah 00:53:43 for what we are doing "static" is maybe a bit misleading 00:54:24 honestly, I don't know how anyone managed to write a web server in python without losing their mind. I mean, it's okay for smaller projects, but at least for me the way the syntax for more complicated class design drives me nuts. I'm thinking about checking out GoLang. 00:54:25 the GameDataHandler class does subclass StaticFileHandler, but it's actually serving version-specific game data for each particular game 00:56:37 so it is backed by static files in some sense, but it's not fully static in the way that those docs mean when they suggest nginx 00:57:04 I saw all the files inside gamedata/UUID... I guess that's what you mean? 00:57:09 yeah 01:08:59 advil, no clue how it worked at all 01:10:02 I remember seeing something about the path validation only being added in 3.1 01:10:41 btw I should have finished testing the chroot stuff by next weekend 01:10:45 there's definitely code that does it in the edlothiel version 01:10:55 I think it changed somehow in 3.1? 01:22:31 yay, gaem back! <3 01:59:42 https://kite.com/python/docs/tornado.web.StaticFileHandler.validate_absolute_path (new in 3.1) 02:00:28 Is this the right spot for crawl.akrasiac.org right? 02:00:50 I have some sort of admin access after the update and I'm pretty sure I shouldn't have that 02:00:55 ConBadger: pretty much, yeah. 02:01:00 https://imgur.com/Eq7PZYP 02:02:52 It was spoken about earlier (the admin panel) because some maintenance/updating was i progress, but yeah doesn't seem like everyone should have it, I'm not one of the admins but I'll check mine to see if it has it as well. 02:03:50 yes - I have it also, so it's not just you. I don't know if it would atually work though if used, maybe there are additional permissions checks at that stage. 02:04:15 For obvious reasons I don't want to test that, I'd rather keep my account 02:04:50 advil was active here last 12:10am (about an hour ago) 02:05:11 I wonder if they forgot to switch a setting back 02:06:51 I'm taking a look 02:07:32 ahh yeah I just found gammafunk's message - a bit after the fact :) [11:03:58 pm] <+gammafunk> wow, this Toggle admin panel button 02:07:32 [11:04:09 pm] <+gammafunk> "Please do not use this frivolously!" 02:07:57 thankfully this is a client rendering issue; the variable for whether to show it at all in client is not working, but it does not actually let you send a message if you're not an admin 02:08:10 would you like me to test it? 02:09:48 I've already tested it 02:16:09 hrm, I see the problem actually 02:16:27 is the message facility actually working? 02:18:33 I think I see 02:28:51 let me test this fix locally, at least 02:35:30 the message facility? if you mean are the notices/messages visible - then yes. First I had in red "Announcement: CAO will go down for maintenance at 11pm EDT on Mar 15, 2020." but that was from awhile ago... currently in green "5, 2020: the planned maintenance should be completed; let us know if you spot any issues." 02:35:49 that's not the message facility 02:36:01 that's static text added to the lobby page 02:36:42 message facility only seen in chat and message log if you're currently playing 02:37:00 ahh I do not know, I wasn't in a game 02:37:07 been sitting in the lobby 02:37:28 I do not recall seeing it before though 02:47:32 03gammafunk02 07* 0.25-a0-611-g6918904: Properly set the admin flag in the WebTiles client 10(4 minutes ago, 1 file, 2+ 1-) 13https://github.com/crawl/crawl/commit/6918904177e8 02:57:22 Unstable branch on crawl.akrasiac.org updated to: 0.25-a0-611-g6918904 (34) 02:58:52 CAO going down for a quick restart 02:59:01 Webtiles server restarted. 02:59:33 gammafunk: you got it 03:01:53 !tell advil Had to fix a quick debug value you left in the webtiles code in 6918904; all users could see admin panel, but it was just client-side visibility and didn't actually work for non-admins. Updated and restarted CAO webtiles. 03:01:54 gammafunk: OK, I'll let advil know. 03:10:38 Windows builds of master branch on crawl.develz.org updated to: 0.25-a0-611-g6918904177 03:11:47 Unstable branch on crawl.beRotato.org updated to: 0.25-a0-611-g6918904 (34) 03:32:45 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-611-g6918904177 (34) 03:48:32 Monster database of master branch on crawl.develz.org updated to: 0.24-a0-443-g80245de385 09:54:58 oh hahaha yes I was actually testing what happens if a non-admin user forces the panel to be visible in the browser 09:54:58 advil: You have 2 messages. Use !messages to read them. 09:55:06 guess that got a trial by fire 10:59:09 when you get the might/agi status, str/dex number goes blue but does not increase 10:59:19 (from DS mutation at least) 11:03:57 if you say open your inventory, does it change? 11:14:06 i'm not sure, i'll check later 11:20:04 03advil02 07* 0.25-a0-612-g9bc587e: Suppress most access logging for tornado >= 3 10(2 minutes ago, 1 file, 10+ 1-) 13https://github.com/crawl/crawl/commit/9bc587e01597 11:21:48 oh hm I wonder if that try will actually succeed no matter what 11:27:38 03advil02 07* 0.25-a0-613-gf4b2665: Tweak the logging fix 10(39 seconds ago, 1 file, 3+ 3-) 13https://github.com/crawl/crawl/commit/f4b266557c28 11:28:56 excited for the day when no one is using tornado 2 11:32:44 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-613-gf4b266557c (34) 11:39:12 Tornado 2? A level 9 Air spell is hard enough... oh, never mind 12:24:25 Stable (0.23) branch on underhound.eu updated to: 0.23.1-93-ge536e68a2c 13:07:32 Unstable branch on crawl.akrasiac.org updated to: 0.25-a0-613-gf4b2665 (34) 14:15:05 Show AC for !lignification 13https://crawl.develz.org/mantis/view.php?id=12224 by jejorda2 14:32:41 advil: do you have any advice for generally how to find out what code is slowing down crawl? I'm trying this patch http://dpaste.com/2MY2VQ3 and whenever I get a screen full of ghosts, the UI becomes very unresponsive 14:32:55 seems to be related to ghosts specifically, and it's even somewhat slow in console 14:33:44 I use a sampling profiler to do stuff like this; the one I use is mac specific (Instruments, part of xcode), but there are others around 14:35:39 ok, looks like linux does have a number of options 14:37:06 aidanh has also done a bunch of crawl profiling at times, iirc in linux 14:51:54 cloud immunity doesn't show in %, maybe it should? since faith is there 15:36:09 Demonspawn monsters are able to use 'Sap Magic' through glass. 13https://crawl.develz.org/mantis/view.php?id=12225 by Yermak 15:42:54 ooh, good bug 15:50:27 -!- aidanh_ is now known as aidanh 15:54:01 more CAO updates: I believe I've gotten gzip precompression for morgue files working; I'm testing it out with tstbtto (whose directory went from around 9G mostly of spammy bot log files to 1.7G). It should be transparent and any .txt link will work as before 15:55:03 the only remaining issue is that I'm not sure how to get the directory listing to be more intuitive (http://crawl.akrasiac.org/rawdata/tstbtto/) -- right now it shows all morgue files as .gz files, but clicking on them opens them as text 15:55:16 (only .txt.gz files are affected by this change) 15:55:41 does anyone know enough about apache to know if there's a way to pretend in the directory listing that those are just .txt files? 16:04:45 as in, you want their extension to be listed as .txt as well? 16:08:00 yeah, if that's possible 16:08:30 I'm seeing some options where I could at least change the icon or something 17:10:11 if you see 2 dancing weapons that are the same, they are noted as "2 dancing weapons" instead of "a +7 dagger, a +7 dagger" 17:10:27 http://prntscr.com/rhf48i 17:54:32 well, the good news is that testing in a normal, optimized non-debug build, ghost zig levels are responsive even when filling the screen with ghosts 17:54:40 guess full debug build slows things down a lot 17:55:10 it definitely does 19:07:42 hrm, the zig permastore has 45 ghosts, yet the code seems to indicate a max of 20 19:14:36 I assume that just means it won't get updated until it expires enough ghosts 19:14:57 ah, looks like it won't ever shrink, just rewrite ghosts over time 19:23:36 Unstable branch on underhound.eu updated to: 0.25-a0-613-gf4b266557c (34) 20:23:50 yeah, many of the distributed permastores exceed the max if it's being filled during regular play 20:23:57 can't remember why I did it that way 20:45:31 what do you mean the levels are responsive? Like they generate quickly? Or does ghost AI use a lot of CPU 20:45:51 add a cloud generator too :) 21:04:05 no, not the levels, just using the UI becomes slow on a full debug build with a los filled with mostly ghosts 21:04:27 level generation is fine, but when you have 15x15 monsters in view and it's mostly ghosts, very unresponsive 21:04:41 seem a bit lagged maybe even in normal build, but still very responsive 21:04:43 is it the same for other ghost_demons? 21:04:51 that's a good question, let me try 21:05:23 can I set the number of zigs completed easilly using the save editor? I guess not 21:05:36 also, sad to report that the precompression trick only makes a difference for a few user accounts like tstbtto 21:06:20 were morgues the major culprit for disk usage on CAO? 21:06:32 it's very hard to answer that because any global stats take an extremely long time to run and lag people currently playing 21:06:47 but I think based on the top user accounts it's ttyrecs, not morgues 21:07:04 tstbtto is (was) the biggest user account, and its morgues were misleading 21:07:43 even qw (second place) isn't so bad for morgues, it is mostly ttyrecs 21:07:59 why do we save ttyrecs eternally again? 21:08:27 heh, for hysterical raisins, just like it says in rax's readme! 21:08:43 seriously, it is pretty neat to be able to watch ttyrecs from old games where people do crazy things 21:08:57 in an ideal world we'd have them around somewhere to look up 21:09:38 obviously none of that should come at the expense of having a worker server 21:10:48 ??tstbtto 21:10:48 I don't have a page labeled tstbtto in my learndb. 21:11:16 I guess one option to buy some time might be to recompress them as xz, does footv support that? 21:11:17 advil, nice work on the server update 21:11:45 thanks, & thanks for your help! 21:11:45 tstbtto is a bot that used to run on cxc 21:11:55 I don't think footv currently support xz 21:12:10 it produced hilarious morgues like this one: http://crawl.akrasiac.org/rawdata/tstbtto/morgue-tstbtto-20160609-031656.txt 21:12:40 what turned out to be the problem in the end? tornado 2.4 did some unexpected path validation? 21:12:51 I never actually figured that out 21:13:07 tornado 2.4 definitely does do path validation, but I didn't change the version of 2.4 on CAO at all 21:13:27 hilarious 21:13:52 I ended up using your trick of just manually setting self.root after the initialize call, I don't know why the old value of "." stopped working with these updates 21:14:19 on old tornado if you pass "/" it gets buggily parsed as "//" 21:14:29 which doesn't match anything 21:14:48 but from the 2.4 path validation code I couldn't see why "." would have ever worked 21:15:31 glad that worked out; updating the other servers should be reasonably straightforward now that we know what to expect 21:15:47 and now that we have a working version in master 21:15:50 yeah, I hope so 21:16:06 don't forget the thrill of 5 people across irc and discord reporting to me that they had admin access on CAO after the restart and after advil had gone to bed! 21:16:24 that did look exciting 21:17:09 despite it all I resisted the urge to tell all of CAO about necromutation using your new feature, so I'm proud of myself 21:17:09 luckily it was only because I had been testing that exact case to make sure that non-admin users could never do anything even if they messed with the dom to show the panel 21:17:15 alexjurkiewicz, did you have a patch for supporting fallback templates? 21:17:23 yeah, it didn't take to long to figure out what you'd done, thankfully 21:17:31 s/to long/too long/ 21:17:44 did you try out the announce thing when doing the shutdown? 21:17:59 that's needed for webserver-dev-env for the new rebuild / save download pages 21:18:01 no, but I did test it locally when checking my fix, and saw that it worked 21:18:22 oh, but I did notice that it's supposed to make a message log (as in the game message log) announcement as well as chat? 21:18:26 and that did not seem to work 21:18:47 or at least it seemed to be announcing in two places 21:18:48 ah, that will also require an update of the install script on all servers 21:19:28 hm, that worked locally for me; it sends the message to the message window as well as chat (does not log it as a note though) 21:19:45 yeah, I didn't see any message window message 21:20:23 advil, btw could you take a glance at the webserver-dev-env python tooling setup some time? I'm not sure whether that's a good fit for your python setup; I tried to keep it as setup-agnostic as possible 21:22:34 sure 21:22:36 gammafunk still works locally for me (should look like this https://www.dropbox.com/s/u5vfb6s2x2m0hr2/Screenshot%202020-03-16%2021.22.13.png?dl=0) 21:22:53 hrm, let me test it again 21:23:18 aidanh: not for non-changes. I was going to do it after merging my current PR 21:23:50 aidanh: in webtiles-changes, I moved the repo templates to banner.html.original, and then load banner.html if it exists preferentially before falling back to .original 21:25:02 advil: I don't see any in-game message; running full debug again, if that makes a difference, and playing as the same user sending the message: https://cdn.discordapp.com/attachments/325651923514556417/689282974419648560/unknown.png 21:25:03 right, we also need to merge that 21:25:06 possibly having it also in chat is overkill, my thinking was that people might miss the in-game thing if they are in a big fight, but that the in-game thing is needed for console players 21:25:43 chat seems a good idea, it gives a new message notification even if they don't have chat open, which might get their attention 21:25:59 guess you could get real fancy with a marquee over the gameplay area, but that's a bit invasive maybe 21:26:11 just to double check you rebuilt the crawl binary right? 21:26:13 aidanh: i've just rebased but it's completely untested. I'll deploy to CPO and let it bake for another week or so I think 21:26:47 oh, maybe that's it 21:26:48 also, is there anything in the log when the announcement is sent? 21:26:49 aidanh: advil: I was planning to run latest and greatest python + tornado on CPO long-term, but would you prefer me to test a different / more common combination? 21:26:59 good call, I see tileweb.cc just rebuilt 21:27:04 heh ok 21:27:16 yeah it'll silently ignore the message probably 21:27:48 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1253 * 0.25-a0-614-g62c4c39: webtiles: Add support for customised env vars 10(10 weeks ago, 3 files, 21+ 3-) 13https://github.com/crawl/crawl/commit/62c4c3927cf8 21:27:48 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1253 * 0.25-a0-615-g1781067: webtiles: Allow loading games via function 10(10 weeks ago, 3 files, 104+ 40-) 13https://github.com/crawl/crawl/commit/1781067247c9 21:27:48 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1253 * 0.25-a0-616-g7dc50c5: webtiles: Add basic requirements.txt file 10(10 weeks ago, 1 file, 2+ 0-) 13https://github.com/crawl/crawl/commit/7dc50c53dba5 21:27:48 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1253 * 0.25-a0-617-g2e5372a: webtiles: Update README 10(10 weeks ago, 2 files, 57+ 49-) 13https://github.com/crawl/crawl/commit/2e5372aa25c8 21:27:48 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1253 * 0.25-a0-618-g0e8a410: webtiles: Allow dynamic reload of games config 10(10 weeks ago, 3 files, 57+ 37-) 13https://github.com/crawl/crawl/commit/0e8a4109c46e 21:27:48 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1253 * 0.25-a0-619-g80c9383: webtiles: Allow specifying -dir and CWD for games 10(10 weeks ago, 4 files, 19+ 1-) 13https://github.com/crawl/crawl/commit/80c938398dd8 21:27:48 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1253 * 0.25-a0-620-g1cd3f65: webtiles: Improve load_games documentation 10(10 weeks ago, 1 file, 20+ 10-) 13https://github.com/crawl/crawl/commit/1cd3f650629f 21:27:48 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1253 * 0.25-a0-621-g8a94496: webtiles: Allow editing blank rc files 10(9 weeks ago, 1 file, 7+ 2-) 13https://github.com/crawl/crawl/commit/8a9449665d34 21:27:48 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1253 * 0.25-a0-622-g859ac3f: webtiles: Allow per-game milestone path config 10(9 weeks ago, 3 files, 34+ 7-) 13https://github.com/crawl/crawl/commit/859ac3f297e7 21:27:48 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1253 * 0.25-a0-623-g0abd74a: webtiles: Always show milestone column 10(9 weeks ago, 1 file, 0+ 4-) 13https://github.com/crawl/crawl/commit/0abd74a8932f 21:27:48 ... and 18 more commits 21:27:59 alexjurkiewicz that's fine with me, if we could get everyone on current versions it would be great 21:27:59 indeed that was it, sorry for false alarm 21:27:59 that's good :-) 21:28:31 is there a reason why we don't use something like setuptools? I'm hardly a python package master, but I use that for my python bots and it works quite well 21:28:45 I kept having the same issue but in browser windows where I needed to force reload the cached js 21:28:45 I've read that requirements.txt stuff is sort of depracated 21:28:55 *deprecated 21:28:59 setuptools is older than requirements.txt :) 21:30:14 what would we use setuptools for? webtiles doesn't get packaged a lot 21:30:59 managing the 'build' and install of webtiles instead of trying to use the makefile? 21:31:39 isn't install done by dgl scripts usually? 21:32:00 from what I'm reading there are some guidelines for the two approaches and they seem to have different use cases 21:32:06 well, we have non-dgl setups 21:32:21 lld, cwz, cpo, for example 21:32:28 alexjurkiewicz as far as I'm concerned the only reasons not to use tornado 6 are (a) bin/str issues that may not be fully worked out, and (b) the timeout issue, which hopefully isn't an issue except on a few servers 21:32:33 ah, no idea how they're set up 21:32:48 alexjurkiewicz so you can find the bin/str issues for me ;-) 21:32:56 I'm not trying to argue for one or the other, more asking 21:33:36 in my own projects I've mainly used setuptools when I'm working on the kind of package that would be distributed on pypi or something 21:33:39 advil: yeah. CPO doesn't have many instances of that timeout triggering. I'm surprised they removed that functionality, maybe it was implemented some hacky way 21:34:08 alexjurkiewicz it's just that asyncio doesn't have support for it in anything like the same way, and tornado 6 forces asyncio 21:34:18 tornado 5 with asyncio also doesn't support it 21:34:53 there's a weird debug setting for asyncio that approximates it, but it didn't look like the kind of thing you really want in a production server to me 21:35:11 not a huge fan of setuptools from my own experience, it's fine for very normal things that you might do with it 21:35:33 A lot of things are sort of saying requirements.txt more for development environments and setuptools for production 21:35:53 but yeah I guess use of one over the other is not as straightforward as I had believed 21:36:04 some system to pin package versions 21:36:42 ...would be nice. You can do that with requirements.txt, but maybe a more modern replacment like pipenv/poetry could be nice 21:44:50 I guess the asyncio way is that anything that can block should be redesigned to be non-blocking 21:45:56 there's stuff like asyncio.wait_for 21:46:43 re: UI slowdown, testing with pan lords level with 670+ monsters, the UI is indeed slow still 21:46:51 let me try again with a non ghost_demon level 21:48:07 yeah, it's slow just an ice monster set 21:48:28 full debug must just be quite slow with so many monsters on a level 21:49:37 Weird: http://crawl.akrasiac.org/scoring/combo-scoreboard.html states that current DS highscore holder is Eld with 44m points, whereas Charly has 68m game. 21:52:06 experimental game? 21:52:09 !hs charly ds 21:52:10 388. Charly the Slayer (L27 DsNe of Cheibriados), escaped with the Orb and 15 runes on 2016-04-24 17:29:04, with 67852526 points after 21171 turns and 9:52:23. 21:52:12 that does look wrong 21:52:15 !hs charly ds x=file 21:52:16 388. [file=cxc/crawl/meta/0.17/logfile] Charly the Slayer (L27 DsNe of Cheibriados), escaped with the Orb and 15 runes on 2016-04-24 17:29:04, with 67852526 points after 21171 turns and 9:52:23. 21:52:22 if it's experimental it wouldn't be on scoring at all, it is 21:52:32 that is a stable version, not experimental, yeah 21:52:45 problems with cxc 0.17 logfile? 21:52:47 there are other ds in the high score list as well 21:52:55 possible after restore of files and rebuild of db? 21:53:03 e.g. http://webzook.net/soup/morgue/trunk/tlatlagkaus/morgue-tlatlagkaus-20180317-031314.txt 21:54:11 and even a couple other charly games 21:55:02 Yermak do you have a github account? if so, https://github.com/crawl/scoring/issues is a place for cao scoring bug reports like that 21:55:14 !hs * ds 21:55:18 1502856. Charly the Slayer (L27 DsNe of Cheibriados), escaped with the Orb and 15 runes on 2016-04-24 17:29:04, with 67852526 points after 21171 turns and 9:52:23. 22:06:59 I opted for requirements.txt with pip tools over setup tools because webtiles isn't installed via pypi, and pip-tools does version pinning 22:08:09 pipenv was an unusably buggy mess the last time I checked, and even poetry had some resolver bugs 22:08:50 you mentioned something else about moving dgamelaunch-config into the main repo? as in crawl/crawl? 22:09:56 alexjurkiewicz, re using latest python+tornado, the eventual plan is to switch to a docker style reproducible environment, but using the chroot scripts 22:10:23 so I'd actually prefer everyone uses that eventually, even if you don't run dgl 22:12:04 gammafunk, yeah, that config has seen significant drift, and some things like the install scripts are not updated with the trunk updates 22:12:58 alexjurkiewicz, I'd be interested to know how you handle updates, since you don't use a chroot 22:13:44 would it be possible/wise to leave the webtiles client stuff in the main repo and move webtiles proper (basically the python side of things and any dgamelaunch-config stuff) to its own repo? 22:14:17 I realize webtiles is its own UI that we support, but it does seem a bit odd to me that we move so much infrastructure stuff into the main repo 22:14:55 but perhaps that's not really that big of a concern 22:15:43 I'd actually prefer things to be more centralised 22:18:01 makes it much easier to keep things in sync, have automated testing, etc 22:25:55 the main problem with the dgamelaunch config repo is that the dgamelaunch format is so verbose and repetitive, which obscures any differences in e.g. the shell scripts 22:26:04 so that really needs some form of templating 22:26:46 another issue is the branch-per-server model, which is horribly unwieldy 22:29:16 if that were just broken out into parallel directory trees in a single master branch, it would be so much more amenable to templating, diff tools, etc etc 22:36:28 advil, re timeouts, isn't tornado behind Apache anyway? 22:36:55 aidanh: what do you mean updates? to the webserver? 22:37:49 i'm all for swapping to running the server in docker. That would make a lot of sense 22:38:57 about the timeouts, I think it's about a timeout where eg the webtiles server is writing new data to the ttyrec file. Until the write is complete communication for the player is blocked -- any commands they say will queue up 22:39:33 oh, sorry, I was getting that confused with idle http timeouts 22:40:26 the update-trunk script will call an install script that's inside the chroot, which will copy the files into place 22:40:53 and in general, I would like to move back to off-the-shelf configuration for webtiles server. But the blocker so far for me has been not wanting to support SSH 22:41:12 because the install script is inside the chroot script, it's not updated automatically, with either the main repo or the dgamelaunch repo 22:41:31 aidanh: yeah, I had sort of started with some templating logic for e.g. dgamelaunch menus and the like so that adding a version/branch could just create the necessary menu files, but there's a lot of stuff to create 22:41:33 I guess for CPO it's less of a concern, because I'm active enough to duplicate the work myself. But I always try to encourage people to use dgl-scripts 22:41:59 none of it is too hard, but there's also the issue of configurability; different servers take somewhat different dgamelaunch menuing approaches 22:42:03 what is dgl-scripts? 22:42:12 he means dgamelaunch-config 22:42:30 the dgl script and all the supporting scripts 22:43:06 ah 22:44:01 that doesn't necessarily require SSH though 22:44:33 I mean, if you don't create the SSH user, not much is gonna happen 22:46:25 right. but it still sets up dgamelaunch and all that. My preference is to not have that at all 22:48:16 how are you running things now? is that all patched out? 22:52:20 gammafunk, re different menuing approaches, yeah, that's a tough problem, especially for servers that support other games as well 22:55:51 aidanh: i don't use dgamelaunch-config at all 22:56:11 aidanh: CPO is running webtiles server only, and a couple of cron jobs to rebuild crawl 22:57:10 https://www.reddit.com/r/dcss/comments/fjuw9s/survey_how_happy_are_you_with_these_changes/ this is interesting to view the poll results 22:57:28 "Trog no longer gifts ranged weapons" has a very mixed response -- roughly equal propotion of players like/neutral/dislike it 22:57:47 whereas spel library is almost universally positive