00:02:20 Unstable branch on CRAWL.XTAHUA.COM updated to: 0.22-a0-533-g4d20840 (34) 00:18:16 !seen Dracunos 00:18:17 I last saw Dracunos at Thu Jun 21 05:32:00 2018 UTC (22h 46m 15s ago) joining the channel. 00:20:19 . 00:20:19 buppy: You have 1 message. Use !messages to read it. 00:20:22 !messages 00:20:23 (1/1) johnstein said (23h 57m 9s ago): experimentals shouldn't be a huge issue so long as there's a stable branch I can point to or some PR in the crawl repo. devil's in the details but generally the only holdup is me carving out time to do it :) 00:27:35 New branch created: pull/792 (1 commit) 13https://github.com/crawl/crawl/pull/792 00:27:35 03pubby02 07https://github.com/crawl/crawl/pull/792 * 0.22-a0-461-gb33feef: New species: Oni 10(3 weeks ago, 23 files, 196+ 19-) 13https://github.com/crawl/crawl/commit/b33feefb8eff 00:27:35 New branch created: pull/793 (1 commit) 13https://github.com/crawl/crawl/pull/793 00:27:35 03pubby02 07https://github.com/crawl/crawl/pull/793 * 0.22-a0-461-ga23b1ca: Smith god: Igni Ipthes 10(2 days ago, 56 files, 1640+ 293-) 13https://github.com/crawl/crawl/commit/a23b1ca415fc 00:27:38 !tell johnstein I will give you my first born son if you can put https://github.com/crawl/crawl/pull/793 and https://github.com/crawl/crawl/pull/792 on cbro as experimentals! 00:27:38 buppy: OK, I'll let johnstein know. 00:27:52 buppy: can you create a branch with all your open changes in it? you get one experimental on cpo 00:28:37 sure but it might break saves 00:33:20 alexjurkiewicz: ok. branch 'buppy' should have everything. https://github.com/pubby/crawl/tree/buppy 00:34:34 :b: 00:35:42 !tell johnstein here's a branch with both features combined: https://github.com/pubby/crawl/tree/buppy 00:35:42 buppy: OK, I'll let johnstein know. 00:35:59 I gotta go 00:36:05 ty 00:36:13 you da man chequers 00:38:28 wow, smith god, talk about a blast from the past 00:38:44 webtiles-changes, smith god, what old fossil will turn up next?! 00:39:20 stalker gonna make a comeback 00:40:15 !tell Dracunos did you ever figure out the problem with epoll module not being found within webtiles chroot 00:40:16 floraline: OK, I'll let dracunos know. 00:52:00 it's so weird looking through 1+ year old irc logs seeing people talk about the exact problems i'm having 01:05:53 interesting, some of my webtiles issues seem to be solved using tornado 2.2.1 which is a bit older than the wiki uses 01:09:18 !seen hellmonk 01:09:19 I last saw hellmonk at Fri Jun 22 05:01:11 2018 UTC (8m 7s ago) quitting, saying 'Quit: Page closed'. 01:09:22 dang 01:27:18 Unstable branch on crawl.develz.org updated to: 0.22-a0-533-g4d20840 (34) 01:39:50 -!- amalloy_ is now known as amalloy 01:42:41 hey 01:42:52 anyone here? 01:47:41 I was informed of this chat by a Drew Hatton, I'm a sys admin who is a huge fan of crawl an I would be willing to donate some time to supporting/optimizing the dcss infrastructure 01:48:20 you can contact me at thompkinwilliaj@gmail.com or by calling 5404292737 01:48:33 hi, that's awesome 01:48:39 i'm new here myself but i know there is much to be improved 01:49:18 that's great, I'm happy to help as needed 01:49:35 i'm currently working through getting a new public server set up 01:49:52 so i am getting very familiar with a lot of the webtiles pain points 01:50:10 what are you hosting it on? 01:50:20 a digitalocean droplet 01:50:36 i know at least one other server is on ec2 and i believe several are on bare metal 01:51:19 and i'm pretty sure everybody's configuration is different 01:52:19 I'm familiar with ec2 and bare metal but have no experience with digital ocean 01:52:56 what issues are you running into? 01:53:06 they're basically trying to compete with some aws services, they try to be completely compatible with AWS apis 01:53:15 and they are just slightly cheaper 01:53:39 well.. lots of issues, and i didn't document them 01:53:39 I play on berotato and that one must be bare metal because I experience lag occasionally 01:53:55 i don't know about that one 01:54:43 like for example the past week i've been trying to get webtiles to run without spinning 100% CPU for several minutes for no reason 01:54:47 and it's been throwing exceptions left and right also 01:54:55 so i'm trying to install different versions of python and tornado 01:54:57 it's all just guesswork 01:54:58 Dossus: Each server admin runs their server independantly, we do have a few things that act centrally, like the Sequell bot (server maintained by a crawl dev who's done most work on the bot) and the CAO scoring web pages 01:55:40 those scoring pages are definitely in need of some work to improve them, although that server does have a couple people doing basic admin already 01:56:05 my problems mostly stem from not being a sysadmin and not having complete or accurate documentation 01:56:15 Windows builds of master branch on crawl.develz.org updated to: 0.22-a0-533-g4d20840 01:56:36 It sort of depends on what kind of work you'd like to do specifically; we have a big need for a single sign-on system that all servers could use, but that's a big project that entails development work 01:56:53 I don't expect to be an admin, I just was hoping to be a resource that could review issues and offer possible solutions 01:57:39 well, we do get people asking technical questions about running servers from time to time 01:57:50 personally from my newbie pov i would be curious to know if most/any of the servers are standardized 01:58:19 and in what ways things could be standardized 01:58:34 and what things are ok to be different. it all looks like a maintenance nightmare to me 01:58:36 Gammafunk: those are the questions I'd be happy to answer 01:58:56 Our recommended server setup is not very streamlined 01:59:11 https://crawl.develz.org/wiki/doku.php?id=dcss:server:setting_up_dgamelaunch_and_webtiles 01:59:12 it's also straight up wrong in some parts 01:59:35 this is a typical "dgamelaunch" setup that many servers use 02:00:39 floraline these are guidelines and basic instructions that admins have to adapt based on their own needs and experience 02:00:46 across multiple platforms setup procedures are sure to be challenging if not outright impossible to follow 02:01:55 are you running Ubuntu 14.04/Debian9 on your digitalocean instance? 02:02:30 yeah i think as someone who is not a sys admin, it's hard to tell where it is OK to deviate from the script 02:02:36 which makes it a little harder to follow a guide like that 02:02:53 not to say it isn't amazing, it's the only way i made it this far 02:02:56 i am running debian 9 02:03:29 right, those instructions are more geared to people who are already admins, they aren't meant to teach you how to be one 02:03:41 yeah, sorry i don't mean to disparage it 02:04:34 if nothing else it's been an excellent learning opportunity so far 02:06:19 that's a great attitude to have, perhaps we could move this conversation to a more dedicated channel and see what we can do about getting this up and running 02:12:22 i don't know any better channel to move it to, this is pretty much the place to be 02:13:09 i do have some questions about one particular issue if you are open to them right now 02:13:18 sure, go ahead 02:14:02 so my webtiles process seems to be going off the rails and taking 100% CPU for a few minutes every few hours 02:14:05 btw I can create a discord channel for help desk so we can have access to voip 02:14:26 to me this seems indicative of a problem and something that would probably not be ok on a live server 02:14:57 how would you figure out what that process is doing during that time, or what caused it? 02:15:35 i used strace and ltrace but all i could glean is that python was trying to throw exceptions very very quickly 02:15:54 I would need to review the logs to see what happens immediately before the cpu hits 100% 02:17:12 with that information alone it is likely a pricing issue. you've likely selected an option that is below the specs needed to run a webtiles server. if it's hitting 100% with no players imagine what will happen when 10-50 people use it at once 02:17:45 that doesn't sound very convincing, it seems odd to me that it would have such clear start and end points 02:18:19 it doesn't appear to correlate with any cron jobs or activity in /var/log 02:18:44 it doesn't feel to me like a performance limitation but rather something is going wrong 02:18:53 anyway we don't have to debug the whole thing i was just wondering how you would approach something like that 02:20:51 i'm having lots of little problems that are just due to my own ignorance, but i'm sure there are some more broad architectural problems that can be solved that are relevant to all servers 02:21:06 I have little to offer with this little information. I'd likely start by googling the fuck out of how to use linux as I'm mainly a windows admin :p 02:21:21 ahh i see :P 02:21:28 that has been my game plan so far and it's working with moderate success 02:23:41 what i really find curious is whether servers ought to be standardized more 02:23:45 it can take you the whole stretch with some patience. it has for me in the past. I'm happy to help as needed 02:24:06 surely it would ease maintenance if everyone is set up in the same or a more similar way 02:24:28 and not only that but it would better enable changes to the code 02:25:10 some changes could be made that would benefit some setups and be detrimental to others, and forking all over the place doesn't sound better 02:25:29 anyway i'm just rambling i really don't know much about this stuff 02:25:34 the server admins are the people you want to talk to 02:26:31 what all their names are, i'm not sure 02:28:51 from my experience changing the code is rare compared to having IT guys figure it out because it's a lot cheaper that way. programmers are rare and expensive 02:29:37 and oftentimes not interested in infrastructure 02:30:36 well you're talking to someone who is a programmer by trade and not remotely a sys admin 02:30:42 for any server admins with issues, I'm happy to offer help, and again you can contact me at thompkinwilliaj@gmail.com or by calling 5404292737 02:31:00 yeah thanks for popping in i bet a lot of people in here would love to chat with you 02:38:46 floraline, do you have logs of what exceptions are being thrown? 02:39:24 i don't 02:39:37 i believe they are tornado exceptions and my log file doesn't seem to print any tornado messages 02:40:22 all i could tell from ltrace is that python was certainly trying to print a traceback 02:40:41 then you should be focusing on finding that information 02:41:38 does tornado have a separate log file? or is it just not writing logs? 02:42:31 as a last resort, you should be able to use strace to dump all written data to a file; you can then grep out any exceptions 02:42:33 i'm not sure. webtiles and tornado both appear to be reporting to some standard python logging mechanism 02:42:43 and that mechanism seems to be intended to print all messages into a log file 02:42:53 i do get some messages but nothing from tornado, i will keep trying to figure that out 02:43:29 actually i did use strace and i didn't get anything that looked like exceptions there 02:43:37 i have a file that i saved of strace and ltrace output 02:44:10 i was able to get tornado exceptions in my stdout earlier when i prevented webtiles from becoming a daemon 02:44:17 i can try that and just leave it running in my terminal 02:44:24 that should yield better results 02:45:34 I believe there's also a log setting in the webserver/ dir that I needed to toggle 02:45:54 that will work; I'd recommend teeing stdout to a file as awell 02:46:54 i'm sort of afraid that if i write it to a file it will fill my hard disk very fast 02:47:12 strace and ltrace at least were just completely spewing text at me 02:47:19 for like 10 minutes 02:47:32 and it was just repeating the same 5ish lines 02:49:35 ok, i've told putty to record my terminal to disk 02:49:56 i have close to 1 tb on this end so i should be able to save it 02:50:28 annoyingly it hasn't happened for a while so i'll just have to be patient. thank you for your advice 02:50:58 I meant teeing stdout to a file, not strace output 02:51:29 most of strace's output will be irrelevant; you could filter it to just write() calls 02:51:35 no yeah i'm talking about stdout 02:51:48 if it's going to spit out exceptions it will probably print a ton of them 02:52:08 it feels like it gets in this very tight loop and throws exceptions as fast as it can 02:54:15 you could poll the file size and kill tornado when it exceeds some limit 02:58:20 Monster database of master branch on crawl.develz.org updated to: 0.22-a0-533-g4d20840 03:06:50 "it feels like it gets in this very tight loop and throws exceptions as fast as it can" makes me laugh. that probably describes a substantial number of python programs 03:12:46 Unstable branch on crawl.beRotato.org updated to: 0.22-a0-533-g4d20840 (34) 03:54:52 -!- amalloy is now known as amalloy_ 04:00:45 floraline, I'm sure you've probably found and changed this by now, but just in case: by default the tornado log file option is commented out in config.py 11:34:56 aidanh: i haven't seen that no, can you point me at it? i dont' see anything about tornado in my config.py 11:47:38 03Muffindrake02 07https://github.com/crawl/crawl/pull/761 * 0.22-a0-421-gf4a58a3: Allow compile by using a semicolon 10(4 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/f4a58a3a24d4 11:57:18 is there a way to run mapstat in parallel or no? 11:59:45 ebering: since the results are independent of each other, you can simply run multiple instances in parallel 11:59:58 you just have to combine the results yourself, but for a simple question that's easy to do 12:00:26 I'm guessing you just care how many instances of a given vault placed out of all levels generated 12:00:48 so you're just at the "Vaults Placed" table (forget the exact name) and the entry for the map(s) in question 12:01:15 cool 12:03:55 oof it's hard-coded to write mapstat.log 12:05:37 maybe a nice project to make mapstat paralellisable and add some automatic combining scripts, but for the question I want to ask it's too much at the moment 12:21:11 floraline: https://github.com/crawl/crawl/blob/4d20840e7c647345606994dbf30781f438e5230c/crawl-ref/source/webserver/config.py#L20 12:26:46 aidanh: oh yes, i did see that. the logging that those options control isn't just tornado, it includes our own webtiles messages too 12:27:09 righto 12:27:10 for some reason my log file only shows webtiles messages and not tornado exceptions 12:27:29 any luck yet collecting those exceptions? 12:27:42 and it's not very consistent, i just switched from tornado 2.1-ish to 2.4-ish and now they are showing up 12:27:55 unfortunately no, i ran webtiles right inthe terminal, not as a daemon this time 12:28:19 and this time the issue did not stop after a few minutes, it stayed at 100% cpu for about six hours 12:28:35 there was absolutely nothing being printed in my stdout 12:29:20 i did find something that looks kind of interesting though, in my /var/log/syslog it says "Starting daily apt upgrade and clean activities" 12:29:51 that's at 6:18:35 and then it says apt-daily-upgrade.timer: adding 20 min 20ish seconds random time 12:29:51 and 6:18:35 plus that amount of time is precisely when the issue began this morning 12:30:04 it could be a coincidence but it's still interesting 12:31:35 that sounds like a rather unlikely coincidence (i.e. quite possibly causally related) 12:32:26 a sampling profiler is also handy in these situations, although unfortunately I don't know if python has great support for that kind of thing 12:34:04 perhaps something like https://stackoverflow.com/a/133384 could help 12:37:29 thank you i'll try this out 12:43:36 some other ideas: strace -c should provide a broader view of what's happening at 100% usage, lsof will tell you what the FDs in strace output correspond to, and there's lots of tools to list current connections, dump traffic, etc 13:07:02 Unstable branch on crawl.akrasiac.org updated to: 0.22-a0-533-g4d20840 (34) 14:02:34 gammafunk: hot new pr coming at you 14:02:36 ghost vaults 14:02:41 cool 14:04:41 New branch created: pull/794 (6 commits) 13https://github.com/crawl/crawl/pull/794 14:04:41 03ebering02 07https://github.com/crawl/crawl/pull/794 * 0.22-a0-534-g9c2f827: Abstract loot scaling for single ghosts 10(3 hours ago, 1 file, 47+ 24-) 13https://github.com/crawl/crawl/commit/9c2f8278aefc 14:04:41 03ebering02 07https://github.com/crawl/crawl/pull/794 * 0.22-a0-535-g240804c: Gozag themed ghost vault 10(16 hours ago, 1 file, 27+ 0-) 13https://github.com/crawl/crawl/commit/240804c7cbea 14:04:41 03ebering02 07https://github.com/crawl/crawl/pull/794 * 0.22-a0-536-g38bfce0: Haunted rock garden ghost vault 10(4 hours ago, 1 file, 22+ 0-) 13https://github.com/crawl/crawl/commit/38bfce04027f 14:04:41 03ebering02 07https://github.com/crawl/crawl/pull/794 * 0.22-a0-537-gce15b12: Reflecting pool ghost vault 10(3 hours ago, 1 file, 18+ 0-) 13https://github.com/crawl/crawl/commit/ce15b12fe1e4 14:04:41 03ebering02 07https://github.com/crawl/crawl/pull/794 * 0.22-a0-538-g7f7a33b: Xom themed ghost vault 10(86 minutes ago, 1 file, 22+ 0-) 13https://github.com/crawl/crawl/commit/7f7a33b9dd56 14:04:41 03ebering02 07https://github.com/crawl/crawl/pull/794 * 0.22-a0-539-g992aa5e: Two sewer themed ghost vaults 10(33 minutes ago, 1 file, 43+ 0-) 13https://github.com/crawl/crawl/commit/992aa5e5f820 14:24:21 wow, commit storm 14:24:27 at this rate I'm not going to have the most ghost vaults 14:28:58 -!- amalloy_ is now known as amalloy 14:48:28 hmm, the veto_fragmentation tag probably shouldn't exist any more right? 14:48:36 also i guess it's fine to allow casting lrd on permarock 14:51:29 Ugh, a semicolon and then a missing type specifier 14:51:31 I shouldn 14:51:39 I shouldn't commit when awake for less than 2 hours 14:52:19 !ghostvaultkills -1 14:52:20 91. mewnfare the Stinger (L5 OpVM), slain by an ooze on D:3 (gammafunk_ghost_arena) on 2018-06-22 18:43:22, with 137 points after 3118 turns and 0:12:13. 14:52:21 !ghostvaultkills -2 14:52:22 90/91. NalNutta the Chopper (L6 DsBe of Trog), mangled by holliwood's ghost on D:3 (biasface_ghost_orc_armoury) on 2018-06-22 18:31:27, with 297 points after 2321 turns and 0:05:55. 14:52:23 !ghostvaultkills -3 14:52:24 89/91. CakeIsGreat the Fighter (L10 GrFi of Okawaru), slain by a ten-headed hydra (kmap: gammafunk_ghost_hydra_chop) on D:7 on 2018-06-22 17:40:05, with 3524 points after 7244 turns and 0:13:29. 14:52:25 !ghostvaultkills -4 14:52:26 88/91. slep the Carver (L12 DsGl of Okawaru), slain by ezero's ghost on Lair:2 (ebering_ghost_inner_flame) on 2018-06-22 17:30:11, with 23947 points after 15796 turns and 1:14:51. 14:52:27 !ghostvaultkills -5 14:52:28 87/91. ThisFiasco the Digger (L6 GrEE), slain by Scuka's ghost (kmap: gammafunk_ghost_necromancy) on D:3 on 2018-06-22 15:04:04, with 412 points after 3668 turns and 0:16:39. 14:52:43 Wow, someone has already died to biasface's vault 14:54:10 I just wonder if someone in depths will ever die to a juggernaut/titan zombie punch 14:54:42 -!- panicbit3 is now known as panicbit 14:55:07 large-scale code architecture question: why are the differences between tiles and webtiles handled through #ifdefs rather than polymorphism? 14:55:40 actually, for "tiles and webtiles" read "console, tiles and webtiles" 14:57:24 03Muffindrake02 07https://github.com/crawl/crawl/pull/761 * 0.22-a0-420-gfa96221: Add lightning-themed vault 10(7 minutes ago, 1 file, 13+ 0-) 13https://github.com/crawl/crawl/commit/fa96221dd9c9 14:57:29 nobody's done the work to change it, nobody knows (but it's possible to guess some of them) what will break if you change the rule that you get only one of them 14:58:13 not sure I can dig up where we ran into one. there are actual code differences between sdl tiles and console/webtiles 14:58:36 which would be hard to handle with polymorphism because they're spread across multiple functions in some cases 14:58:57 "It ain't broke, don't fix it"? 14:59:18 "it ain't broke, nobody got time" 14:59:29 MarvinPA_: does that fragmentation veto affect shatter? 14:59:50 the background for the question is I'm trying to do something that involves looking at all three codepaths 15:01:05 and I suspect that if it was done the other way I'd need a lot less time to understand the code (because they diverge in a whole lot of places and it's often not obvious why) 15:01:23 "nobody got time" is a very good reason, of course 15:01:48 you migth well need more, because you'd be faced with things done in one function in some cases and in another function in others 15:02:04 and refactoring isn't really an option unless you want to completely re-architect the game loop 15:02:44 this is not a matter of "oh just change a few things", you will literally need to redesign the entire game to change it 15:02:44 it has been poked at before 15:02:55 (and a large reason for that is that it started out as console only, sdl and webtiles are both later hack-ins) 15:03:51 (hcked in because, again, complete redesign of the game needed to do it "right") 15:04:49 the historical explanation is pretty much what I expected, yeah 15:05:44 as an answer. but there might have been a different one, so I figured I'd ask 15:06:24 and, yeah, I'm pretty far away from either intending to redesign it myself or expecting it to be redesigned. 15:06:55 The correct answer is that console is the one true way to play this game, all other are inferior 15:09:46 neofelis: What specific crazy divergence are you referring to? 15:10:44 gammafunk: no, there's seperate tags for shatter/disint/lrd 15:11:02 (and disint should be changed to dig now i guess, but maybe that requires some save compat) 15:12:19 oh, interesting 15:13:13 03ebering02 07https://github.com/crawl/crawl/pull/794 * 0.22-a0-540-g0e5b7b0: Scale Gozag ghost vault gold 10(7 minutes ago, 1 file, 16+ 1-) 13https://github.com/crawl/crawl/commit/0e5b7b0e4cc9 15:13:13 03ebering02 07https://github.com/crawl/crawl/pull/794 * 0.22-a0-541-gf3eec90: fixup! Xom themed ghost vault 10(0 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/f3eec902a02a 15:13:27 fixup! 15:16:32 I forgot chaos brand!! 15:16:44 muffindrake, "crazy" is def. not the word for it. but the example that made me ask the question was that for console, viewwindow() redraws things, and for tiles not 15:22:11 The path of pain: a vault with 10 chaos arbalests 15:22:46 for tiles, redraws happen while waiting for player input 15:23:32 so if you're in the targeter, it will redraw things when you move the cursor in either mode, but it will redraw them at a different time depending on the mode 15:23:57 and maybe update things differently and at different times, though I'm still not clear on how that part works 15:25:21 tiles has an active agent somewhere else (webtiles has a client running in the browser, local times has sdl event loop) and supports multiple windows. console does neither. and webtiles relies on this because it uses a "captive" console which should not behave like a game UI to make ttyrecs 15:26:15 as such, console also does extra refreshes because ttyrecs need there to be regular full redraws so it can find a good place to start playback 15:28:45 What degree of murderous may a vault be? Very? Almost certain? Fair? Unfair? 15:29:05 Does it depend on the depth? 15:30:55 you can have a murderous vault as long as you provide fair warning. like the oka arena where there are glass walls around it and you can see the nasties waiting. there's another one I hit in lair recently where the monsters get nastier as you go deeper, giving you a chance to think about whether you want to turn those next couple corners to find the next one 15:32:15 i think it's also reasonable to measure it in like, what percentage of the player's resources will they have to expend to survive. on D:2, a vault that you can survive with a potion is a big deal, but in depths people often spend multiple potions, scrolls, and wand charges 15:33:36 I'd say "less lethal than some transporter vaults" but that covers basically anything 16:36:40 hmm maybe veto_fragmentation should still be applied in linesprint, is lrding all of linesprint too op 16:40:40 haha 16:41:28 does permarock not allow LRD? Maybe that was removed 16:43:40 you can't lrd permarock yeah 16:45:07 from glancing at git history it looks like at some point you could cast it on permarock (but not destroy it), but that changed at some point between 2013 and now 16:47:40 or hmm, maybe not actually. but either way i was going to enable casting it on permarock now 16:51:19 ebering: jboons just posted about how a ghost in your immolation vault stood in flame clouds until it died; maybe it would be better to just replace those with black smoke fog machines? 16:52:22 the problem for ghosts is that they have pretty high hp and are likely to think it's safe to step in the cloud, then they never decide to step out of it 16:53:26 yeah 16:53:57 I am very familiar with monster cloud ai. Black smoke would probably still get the point across 16:54:29 another possibility is not not place fire on the 3 tiles closest to the door 16:54:35 so that if the ghost is tracking the player it stays safe 16:54:49 yeah, you could try placing it in the back of the vault 16:55:15 however I think chunks and black smoke still get the idea across and will guarantee no monster deaths 16:55:50 in one test I did, a manticore died out of los as I rested 16:56:10 do you want a pr or do you just want to just make the change 16:56:35 you can make the pr if you have time, if not, no worries, I'll get to it in my next push 16:57:19 !lg * kmap~~ebering 16:57:20 7. slep the Carver (L12 DsGl of Okawaru), slain by ezero's ghost on Lair:2 (ebering_ghost_inner_flame) on 2018-06-22 17:30:11, with 23947 points after 15796 turns and 1:14:51. 17:03:58 Qazlal worshippers get messages when donning or doffing cloud immunity 13https://crawl.develz.org/mantis/view.php?id=11500 by damerell 17:14:28 New branch created: pull/795 (1 commit) 13https://github.com/crawl/crawl/pull/795 17:14:28 03ebering02 07https://github.com/crawl/crawl/pull/795 * 0.22-a0-534-gbbc7872: Remove flame clouds from inner flame vault (jboons) 10(5 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/bbc78720943b 17:29:03 03ebering02 07* 0.22-a0-534-gbbc7872: Remove flame clouds from inner flame vault (jboons) 10(19 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/bbc78720943b 17:29:03 03gammafunk02 {GitHub} 07* 0.22-a0-535-gbbb2ed9: Merge pull request #795 from ebering/tweak-inner-flame-vault-pr 10(7 seconds ago, 0 files, 0+ 0-) 13https://github.com/crawl/crawl/commit/bbb2ed93a3c7 17:50:30 03MarvinPA02 07* 0.22-a0-536-g9b9f9d0: Remove an unused param 10(78 minutes ago, 3 files, 4+ 13-) 13https://github.com/crawl/crawl/commit/9b9f9d0a2d30 17:50:30 03MarvinPA02 07* 0.22-a0-537-ge9e9f76: Remove veto_fragmentation, replace veto_disintegrate with veto_dig 10(60 minutes ago, 12 files, 33+ 37-) 13https://github.com/crawl/crawl/commit/e9e9f763428c 18:10:02 Unstable branch on crawl.jorgrun.rocks updated to: 0.22-a0-537-ge9e9f76342 (34) 18:20:26 Unstable branch on underhound.eu updated to: 0.22-a0-537-ge9e9f76342 (34) 18:38:05 03pubby02 07https://github.com/crawl/crawl/pull/793 * 0.22-a0-462-g473353f: Prevent igni worshipers from wearing non-crafted artefacts 10(11 minutes ago, 9 files, 61+ 1-) 13https://github.com/crawl/crawl/commit/473353f7ac22 18:38:05 03pubby02 07https://github.com/crawl/crawl/pull/793 * 0.22-a0-463-g9ebc612: Fix useless check in igni artefactize abilities 10(3 minutes ago, 1 file, 6+ 23-) 13https://github.com/crawl/crawl/commit/9ebc612624ec 19:18:50 The build has errored. (master - e9e9f76 #9830 : Chris Campbell): https://travis-ci.org/crawl/crawl/builds/395650552 19:20:32 neofelis: the changes in the ui-overhaul branch are intended to unify a lot of what you describe as "updating things differently and at different times" 19:22:56 among other things, it replaces ad hoc writing to the console with an intermediary widget layer, which handles all rendering/layout 19:23:13 that's great! 19:25:37 I mean, again, it's not bad code currently. but introducing a rendering layer for console is certainly an improvement 19:40:40 -!- eb__ is now known as eb 19:42:43 it's not bad, unless you want to make changes to it 19:43:04 the same rendering layer is used for sdl tiles as well, which fixes all kinds of weird issues 19:44:06 and webtiles now uses html for nearly everything, rather than emulating a terminal emulator 19:47:02 of course, the primary function of this branch is to trounce gammafunk's commit count 19:56:52 gammafunk: that lrd tree thing is definitely a bug in cell_see_cell or how trees are defined 19:57:13 yeah 19:57:14 los-type.h's comments are pretty clear that trees should be opaque to LOS_NO_TRANS 19:57:39 that's what I figured, that it was simply at the level of cell_see_cell and they somehow didn't get the right opacity 19:58:02 but I haven't looked at the code 19:59:16 holdover from when trees had that opaque-when-depth-two thing? 20:00:08 aidanh: sadly you're already way ahead for this release http://dpaste.com/3FQDXXE 20:00:21 I doubt I'll even catch up to advil 20:02:44 hmm, how did you calculate that? 20:03:25 git log --use-mailmap --pretty="%aN" 0.22-a0..HEAD | sort | uniq -c | sort -bgr | head 20:04:20 I may not have the absolute latest trunk but it's pretty recent 20:08:44 I think I have wondered the same thing before, but I wonder why github's graphs are different 20:09:18 https://github.com/crawl/crawl/graphs/contributors?from=2017-12-23&to=2018-06-23&type=c 20:09:43 maybe something about how it handles PR merging 20:09:44 oh, actually 20:10:34 I wasn't using --no-merges, but when I did, it still doesn't agree 20:11:51 advil: ah, might it be using author date? 20:12:16 doesnty commits doesn't look right 20:13:15 but I should use --no-merges in my command probably, regardless 20:13:47 %git 0.22-a0 20:13:47 07gammafunk02 * 0.22-a0: Preliminary debian changelog entries for 0.21 and 0.22 10(6 months ago, 1 file, 10+ 0-) 13https://github.com/crawl/crawl/commit/fd72fac9dab4 20:16:14 that's probably a factor, when I use --since instead of the tag my numbers get closer to matching (not aidanh's though) 20:17:27 I think aidanh merged a bunch of commits of his to trunk that he had authored in the 0.21 cycle 20:23:08 -!- yesno_ is now known as yesno 20:24:06 I did, yeah; menu related stuff 20:31:22 i didn't expect to be as high as fifth 20:33:23 you had a bunch from the spell library that I can remember 20:34:24 guess we got a TAG_MINOR bump so ghost bones/permastore files will be back to square one 20:35:05 just like my ccache is back to square one right now ;_; 20:38:16 hm I wonder what will actually happen 20:38:27 that might just break the permastore unless someone manually clears it 20:38:41 since it only uses one file 20:40:37 though maybe any problems will be had by players who *don't* upgrade 20:41:53 I'd be higher if my prs ever merged 20:42:01 ;P 21:01:41 gammafunk: doing some digging it seems that bug is not in cell_see_cell 21:01:52 but beam.cc 21:02:03 hm no 21:02:05 augh 21:05:44 yes 21:12:49 ok 21:12:59 !source beam.cc 6021 21:13:00 https://github.com/crawl/crawl/blob/master/crawl-ref/source/beam.cc#L6021 21:13:04 there is the offending conditional 21:13:11 the problem is that trees are not FFT_WALL 21:16:16 well I wonder if that should just check a more general function looking for solid features 21:16:16 instead of checking wall or door 21:16:49 but yeah I don't know how determine_affected_cells proceeds thorughout the lifetime of a bolt explosion 21:18:01 I suppose you wouldn't want just solid feature since statue actually should allow the cells to be effected 21:18:47 but it feels a bit odd, this conditional and directly adding a new condition might make it weirder 21:19:21 maybe we can just make trees be considered FFT_WALL with no negative repercusion though 21:19:24 it recurses depth-first adding counts to the explosion_map& m for how many times each cell was hit until it stops 21:19:45 and that conditional is part of determining stops when stop_at_wall is true 21:23:31 looks like it's more correct to say it's determining when at_wall is true, or rather when stop_at_wall should be actually be considered 21:24:08 it feels like a bit of a wacky conditional but it does look like addition the condition there would fix the problem 21:24:37 I'd wonder if there's not a more general...are statues also considered wall? 21:24:42 !source feat_is_wall 21:24:42 1/1. https://github.com/crawl/crawl/blob/master/crawl-ref/source/terrain.cc#L74 21:25:09 statues are not wall 21:26:49 only things you think should be are wall 21:26:49 statues are solid though 21:29:23 !source feat_is_solid 21:29:24 1/1. https://github.com/crawl/crawl/blob/master/crawl-ref/source/terrain.cc#L366 21:29:45 (I'd be wary of telling it trees are FFT_WALL: what does passwall do then?) 21:32:05 yeah, I assumed it might have other side effects 21:32:45 ebering: looks like the comment mentioning statues is a bit erroneous, since statues aren't relevant to the case it's documenting, but I see what's happening there 21:33:37 rock wall (or stone wall or any other kind of actuall wall or closed door) are relevant; it's saying don't consider the explosion blocked when checking the center of the explosion 21:36:28 I think passwall would be fine in that it only allows rock in the first place 21:37:56 it does sort of seem that trees should be considered wall but I'm not sure if there would be other problems with that 21:42:37 yeah, tracking down what making trees FFT_WALL does doesn't seem too bad 21:44:33 might do weird things with bloodspatter? 21:47:33 oh it's used all over dgn-layout via a wrapper function 21:48:39 also walls don't get floor tiles 21:48:48 oh, right 21:48:50 my suggestion: trees should not be FFT_WALL 21:48:55 and trees need floor tiles 21:50:25 seems like they probably should be but there should be some better logic available to discriminate for cosmetic purposes, but yeah 21:50:36 maybe adding a condition in beam.cc with a comment is best for now 21:51:01 well there is feat_is_tree 21:51:03 FFT_INORGANIC_WALL??? 21:51:38 which it seems the tile code uses proactively to tile them correctly? 21:52:00 it seems like scattered throughout the codebase is an assumption that trees and walls are different even though they often aren't 21:52:13 except when they are 21:52:16 right, that's sort of what I was getting at 21:52:47 it used to be the case that you needed two tiles worth of trees to break los, so trees were definitely more different from other walls in the past 22:03:38 New branch created: pull/796 (2 commits) 13https://github.com/crawl/crawl/pull/796 22:03:39 03ebering02 07https://github.com/crawl/crawl/pull/796 * 0.22-a0-538-gc63589d: Don't treat trees as half-opaque for exclusions 10(59 minutes ago, 1 file, 1+ 2-) 13https://github.com/crawl/crawl/commit/c63589d8fd15 22:03:39 03ebering02 07https://github.com/crawl/crawl/pull/796 * 0.22-a0-539-ga1b6296: Have trees to block explosions (Reeces) 10(6 minutes ago, 1 file, 5+ 2-) 13https://github.com/crawl/crawl/commit/a1b62967d210 22:44:30 03gammafunk02 07* 0.22-a0-538-g76cc104: And icy themed ghost vault for Dungeon, Lair, and Depths (Biasface) 10(2 hours ago, 1 file, 67+ 0-) 13https://github.com/crawl/crawl/commit/76cc10456c96 22:46:07 Hello there! 22:46:25 So... are there any cleaving weapons with a reach of greater than 1? 22:47:31 I would like to try to account for all reach possibilities, if possible. 22:51:07 I've been trying to resolve a bug: 22:51:09 https://github.com/crawl/crawl/issues/718 22:51:32 I believe I have the code in place that will do this. 22:52:14 I am still in the process of figuring out how to test this. 22:52:21 xom can grant cleaving to players with a polearm 22:52:38 at least I think that effect is still in the game 22:52:59 That implies that maximum possible range is 2. 22:53:22 no, it just gives cleaving to adjacent melee I think 22:53:51 Alright... I think this code will cover that... 22:54:05 I still need to test this under the conditions described in the bug report. 22:54:40 heh, found a bug: in webtiles, when targeting, ':' does not actually hide the beam 22:59:41 imo the correct solution is to remove the ':' command; if you disagree, the fix is to change directn.cc:1133 to #ifndef USE_TILE, and not worth a pr 23:09:58 Unstable branch on crawl.jorgrun.rocks updated to: 0.22-a0-538-g76cc10456c (34) 23:16:21 When you make changes to the code; how do you test it? 23:18:53 you compile it 23:18:56 ??install 23:18:56 install[1/1]: See the following for installation and compilation: https://github.com/crawl/crawl/blob/master/crawl-ref/INSTALL.txt 23:19:09 that document has sections on how to compile the code under various OS 23:20:20 I think the question is more "how do you get the game conditions you want", to which the answer is "mostly wizmode, sometimes also waiting" 23:20:41 (I think Alcon can compile; I'm not sure they know what wizmode is) 23:20:45 I already have it compiled and running. And "Yes", neofelis has the gist of what I was trying to ask. 23:21:10 okay; wizmode is what you get by pressing & in-game 23:21:18 then you can press &? for a list of wizmode commands 23:21:37 you can create monsters, learn spells, set experience level & religion & so on and so forth 23:21:41 I have wizardmode running right now; but I need to create the conditions where I have followers and I need to create a hostile monster with a distortion weapon. 23:22:23 nm... player with distortion weapon... 23:22:28 ... ah... I can do that 23:23:50 &m orc ; hand axe 23:24:02 to create something that has a cleaving weapon :) 23:24:07 Thank you. 23:24:08 (@ Alcon) 23:37:52 After summoning a spriggan berserker wielding a hand axe, whilst, myself, wielding a longsword of distortion... triggered a riposte after a cleave... and the game hasn't crashed, yet. 23:38:11 Is there a more rigorous way I could test this before making a 'Pull Request'? 23:41:25 Arena: 40 orc; hand axe v 40 orc; long sword ego:distortion 23:41:46 that would probably do the trick 23:42:01 Done... no crash from assert. 23:42:42 -!- neofelis_ is now known as neofelis 23:42:55 does that crash without your changes? 23:46:37 Good question... 23:46:40 Let me try... 23:46:57 The build has errored. (master - 76cc104 #9832 : gammafunk): https://travis-ci.org/crawl/crawl/builds/395707283 23:48:27 I have to recompile the whole codebase. :P This could take 10 minutes before I can run the test. 23:51:00 if you're planning on doing much dev work, I highly recommend setting up ccache 23:51:38 Yeah... it has to do with the nature of the command I'm using... 23:52:17 make APPLE_GCC=y NO_PKGCONFIG=y CONTRIB_SDL=y TILES=y 23:53:11 03pubby02 07https://github.com/crawl/crawl/pull/793 * 0.22-a0-464-gccd645a: Fix artefactize prices 10(9 minutes ago, 1 file, 65+ 67-) 13https://github.com/crawl/crawl/commit/ccd645a7e87a 23:53:11 03pubby02 07https://github.com/crawl/crawl/pull/793 * 0.22-a0-465-g4d7a21c: Mark artefacts as useless under Igni 10(7 minutes ago, 1 file, 3+ 0-) 13https://github.com/crawl/crawl/commit/4d7a21cd03b0 23:53:14 if you change those parameters it'll require a rebuild, yeah 23:55:08 buppy =pubby? 23:57:06 I'm seeing that the arena battle sometimes seems to halt early... but I was expecting to see a more serious crash-like behavior. 23:57:52 ... although... 23:58:10 Can monsters trigger their own riposte's? 23:58:30 Igni. There's a name I've not seen in a while 23:59:09 yea I have all craw github PRs on notifications and was kinda stoked to see it make a comeback 23:59:34 I came in to see if it was ready for an experimental and saw that buppy already asked me for it