00:06:30 ^status 00:06:30 24 Crawlers. CBRO disk usage=94% (135GB) | RAM usage=43% (4GB)| uptime/CPU= 00:06:24 up 136 days, 21:38, 1 user, load average: 0.66, 0.70, 0.59 (4 Cores) http://status.berotato.org 00:59:51 big fan of new webtiles server announce functionality, thanks for adding it. Just used it to announce a server restart 01:07:54 there was one issue with game definition loading after USR1 signal -- I've pushed a fix to my PR branch. I think it's ready for merge -- or further cherry-picks. Whichever you prefer advil / aidanh 01:08:58 btw, the crashes I see during shutdown generate crash milestones with reason "Crash caused by signal #15:: Terminated" 01:09:07 does crawl expect some other signal to quit? 01:09:41 I think it uses SIGHUP for clean shutdown? 01:11:31 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1253 * 0.25-a0-738-ga442df4: Enhancements to webtiles game definitions 10(9 days ago, 7 files, 395+ 53-) 13https://github.com/crawl/crawl/commit/a442df417b71 01:11:31 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1253 * 0.25-a0-739-g1265078: Reduce log spam 10(9 days ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/1265078c4fb5 01:11:31 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1253 * 0.25-a0-740-ge2e2f79: Convert README to markdown and update 10(9 days ago, 2 files, 87+ 49-) 13https://github.com/crawl/crawl/commit/e2e2f792ec1e 01:11:31 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1253 * 0.25-a0-741-g6a14cd8: Remove version pin from Tornado dependency (advil) 10(9 days ago, 5 files, 21+ 18-) 13https://github.com/crawl/crawl/commit/6a14cd85a159 01:11:31 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1253 * 0.25-a0-742-g535ec1e: Improve some webtiles log messages 10(9 days ago, 3 files, 9+ 6-) 13https://github.com/crawl/crawl/commit/535ec1ed4696 01:11:31 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1253 * 0.25-a0-743-gc11e514: Add some missing type information 10(9 days ago, 2 files, 7+ 5-) 13https://github.com/crawl/crawl/commit/c11e5143f873 01:11:31 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1253 * 0.25-a0-744-g3ca2e41: Allow overriding some HTML templates 10(5 days ago, 7 files, 12+ 8-) 13https://github.com/crawl/crawl/commit/3ca2e41cf145 01:11:31 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1253 * 0.25-a0-745-g890ea29: Always show milestone column 10(5 days ago, 1 file, 0+ 4-) 13https://github.com/crawl/crawl/commit/890ea29f115c 01:11:31 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1253 * 0.25-a0-746-g96ee691: Fix humanise_bytes unit 10(5 days ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/96ee691d3679 01:11:43 specifically, if the binary doesn't get SIGHUP, it probably doesn't save 01:15:39 yeah, specifically 01:15:51 !source libutil.cc:648 01:15:51 https://github.com/crawl/crawl/blob/master/crawl-ref/source/libutil.cc#L648 01:15:54 alexjurkiewicz: ^ 01:16:33 use of 15 is like using unix kill with default signal 01:17:09 if you use SIGHUP, crawl will abort input, go back to main loop, save, then exit 01:17:09 so that's the clean way of shutting down a process 01:17:21 should also be what's used if websocket or ssh connection is lost 01:17:32 so breaking your connection doesn't cause a progress-losing crash 01:19:58 seems like webserver shut down process sends the wrong signal then 01:20:44 well, shutdown might have to in order to guarantee that the process ends 01:21:18 I suppose it's not a guarantee unless you use SIGKILL, but I don't know much about signals 01:22:35 it looks like the code is meant to send sighup, wait 10secs and then send sigkill if the process still exists 01:22:40 but there is a bug in the timeout code which means sigterm is sent immediately 01:22:48 I wonder if this is due to changed tornado behaviour 01:23:09 actually, that's wrong 01:23:18 yeah, just found that code 01:25:19 in my log, I see output like this: 01:25:21 Apr 12 05:00:03 ip-172-26-2-230 systemd[1]: Stopping DCSS Webtiles Server... 01:25:23 Apr 12 05:00:03 ip-172-26-2-230 python[2271]: INFO: Received signal 15, shutting down. 01:25:25 Apr 12 05:00:03 ip-172-26-2-230 python[2271]: INFO: #0 Shutting down user None id 0 01:25:27 Apr 12 05:00:03 ip-172-26-2-230 python[2271]: INFO: #2440 Shutting down user Buugeng id 2440 01:25:32 [...] 01:25:34 Apr 12 05:00:03 ip-172-26-2-230 python[2271]: INFO: #2517 Shutting down user Odyssey id 2517 01:25:36 Apr 12 05:00:03 ip-172-26-2-230 python[2271]: INFO: #2392 Shutting down user bogs101 id 2392 01:25:38 Apr 12 05:00:03 ip-172-26-2-230 python[2271]: INFO: #2517 P2714 ERR: Crash caused by signal #15: Terminated 01:25:40 Apr 12 05:00:03 ip-172-26-2-230 python[2271]: INFO: #2517 P2714 ERR: We crashed! This is likely due to a bug in Crawl. 01:25:50 then every game crashes from TERM, without any delay 01:26:49 hm. I wonder if systemd is doing this. Sending TERM to every process in the process group 01:28:46 aha yeah. By default 'systemd stop ' sends TERM to all processes in the cgroup 01:32:16 KillMode=mixed in the systemd service definition will send the initial TERM only to the main process (webtiles server) and KILL every process after the systemd timeout is reached (90s) 01:33:27 Anyone know why doesn't DCSS shut down cleanly on SIGTERM? To be, it's a standard "stop cleanly" signal 01:45:47 not sure about that one 01:46:10 do most unix applications actually save if you kill them that way? 02:06:03 I'm most used to SIGHUP meaning "reload configuration from disk" or "restart" and SIGTERM meaning "gracefully terminate" 02:06:22 my definition of graceful termination would definitely include saving any data 02:07:18 probably whoever implemented that signal handling wasn't aware of convention 02:43:14 well, it's a sort of old school shell convention, HUP meaning "the remote terminal has disconnected, so you can't ask for user input and should shut down now" 02:44:06 DCSS probably gets HUPped from dgamelaunch, in other words. But I'll submit a patch to treat TERM the same 02:45:03 Monster database of master branch on crawl.develz.org updated to: 0.25-a0-651-g8c222c7a2b 02:55:51 !tell advil That korean user who had compilation fail reports "My windows version is 8.1, and I've installed MSYS2 from scratch." 02:55:51 gammafunk: OK, I'll let advil know. 03:16:32 New branch created: pull/1347 (1 commit) 13https://github.com/crawl/crawl/pull/1347 03:16:32 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1347 * 0.25-a0-738-g2df5439: Shut down cleanly on SIGTERM 10(12 minutes ago, 1 file, 3+ 0-) 13https://github.com/crawl/crawl/commit/2df543981b05 03:21:59 New branch created: pull/1348 (1 commit) 13https://github.com/crawl/crawl/pull/1348 03:21:59 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1348 * 0.25-a0-738-gfcabaf8: Always save macros 10(9 minutes ago, 2 files, 2+ 7-) 13https://github.com/crawl/crawl/commit/fcabaf8530a1 03:41:58 Stable (0.24) branch on crawl.kelbi.org updated to: 0.24.1-4-ge52af88ad5 03:42:54 Fork (bcrawl) on crawl.kelbi.org updated to: 0.23-a0-3001-g9d8a61a446 03:44:13 Fork (bcadrencrawl) on crawl.kelbi.org updated to: 0.22.1-2575-g1deda5d986 03:57:08 03Alex Jurkiewicz02 {Aidan Holm} 07* 0.25-a0-738-g4d5fadd: Shut down cleanly on SIGTERM 10(53 minutes ago, 1 file, 3+ 0-) 13https://github.com/crawl/crawl/commit/4d5fadd62c62 04:00:02 03Alex Jurkiewicz02 {Aidan Holm} 07[test-make-lint] * 0.25-a0-739-ge4527e4: Fix humanise_bytes unit 10(5 days ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/e4527e4d3b14 04:00:02 03Aidan Holm02 07[test-make-lint] * 0.25-a0-740-gc9e5e28: Fix lint errors in util.py, and add type hints 10(6 days ago, 3 files, 57+ 39-) 13https://github.com/crawl/crawl/commit/c9e5e28b46ae 04:11:35 ha ha aidanh that broke the unit tests, I think 04:11:46 ah just in your branch, misread 04:14:32 :) 05:07:08 New branch created: fix-messaging-crash (2 commits) 13https://github.com/crawl/crawl/tree/fix-messaging-crash 05:07:08 03Alex Jurkiewicz02 {Aidan Holm} 07[fix-messaging-crash] * 0.25-a0-739-ge4527e4: Fix humanise_bytes unit 10(5 days ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/e4527e4d3b14 05:07:08 03Aidan Holm02 07[fix-messaging-crash] * 0.25-a0-740-g67ea76e: Work around crash caused by early-game messaging 10(7 minutes ago, 1 file, 7+ 0-) 13https://github.com/crawl/crawl/commit/67ea76ea5cb3 05:11:56 03Alex Jurkiewicz02 {Aidan Holm} 07* 0.25-a0-739-g5222f23: Fix humanise_bytes unit 10(5 days ago, 2 files, 4+ 4-) 13https://github.com/crawl/crawl/commit/5222f2338962 05:12:21 .. probably should have given that one a better commit message 05:13:25 03Aidan Holm02 07[test-make-lint] * 0.25-a0-740-g9d32e06: Fix lint errors in util.py, and add type hints 10(6 days ago, 3 files, 58+ 40-) 13https://github.com/crawl/crawl/commit/9d32e0615e06 05:18:05 alexjurkiewicz: is there any more info on tornado 3+ support, especially tornado 6? 05:18:29 I remember there was a bug for something with tornado 6, not sure if I missed some new developments there 05:50:52 03Aidan Holm02 07* 0.25-a0-740-g2f0c327: Remove needless conditional compilation for MSVC 10(2 minutes ago, 2 files, 0+ 23-) 13https://github.com/crawl/crawl/commit/2f0c327db6a4 06:19:43 -!- DarwinElf is now known as DarwinFreyson 06:23:07 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-740-g2f0c327db6 (34) 07:17:17 -!- Amnesiac_ is now known as Amnesiac 07:24:14 aidanh: i don't think so. It's running fine with tornado 6 on cpo atm 07:26:10 03Aidan Holm02 07* 0.25-a0-741-g01bdca3: Fix lint errors in util.py, and add type hints 10(6 days ago, 3 files, 58+ 40-) 13https://github.com/crawl/crawl/commit/01bdca394dc3 07:26:10 03Aidan Holm02 07* 0.25-a0-742-gbe65d2b: Work around crash caused by early-game messaging 10(2 hours ago, 1 file, 7+ 0-) 13https://github.com/crawl/crawl/commit/be65d2b916ec 07:32:20 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-742-gbe65d2b916 (34) 07:55:58 alexjurkiewicz: hmm, it appears to be something with my chroot scripts 08:30:36 well, I guess the explains why only alexjurkiewicz has been seeing crashes on restart 08:30:36 advil: You have 1 message. Use !messages to read it. 08:30:51 weirdly, the dgl /etc/init.d/webtiles script sends ABRT? 08:37:10 FYI webtiles server uses SIGTERM for force shutdowns, so that commit probably broke that 08:39:04 as in the webtiles init script? 08:39:29 !source process_handler.py:487 08:39:30 https://github.com/crawl/crawl/blob/master/crawl-ref/source/webserver/process_handler.py#L487 08:39:41 the stale process handler 08:41:30 ...bugger 08:52:58 03Aidan Holm02 07* 0.25-a0-743-g84dc206: Revert "Shut down cleanly on SIGTERM" 10(3 minutes ago, 1 file, 0+ 3-) 13https://github.com/crawl/crawl/commit/84dc20635f85 08:55:21 alexjurkiewicz: had to revert ^ so your .service file will need tweaking 08:59:48 I guess this is why the slightly odd arrangement of signals...not sure if there's a good alternative signal to use for that case 09:00:45 maybe SIGQUIT? 09:01:15 I agree that conventionally TERM leads to the expectation of a clean exit if possible, so it is kind of weird 09:02:32 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-743-g84dc20635f (34) 09:03:12 is there any meaningful difference between an uncaught SIGTERM and SIGKILL? 09:04:33 SIGKILL ime doesn't trigger any of crawl's crash code 09:04:54 whereas SIGTERM does 09:05:08 also alexjurkiewicz why are you running killall instead of sending a signal to webtiles server? 09:05:25 (somewhat orthogonal to the quesiton of which signal) 09:05:48 i think it was systemd that was doing that 09:07:55 I guess we would probably want a backtrace if a process was stalled after SIGHUP, but then again that'd take much longer than 10 seconds 09:10:04 oh, I see...that's probably also not what you want systemd to do but I guess he probably is sorting that out 09:11:52 alexjurkiewicz even without dgl, you could still consider adapting this dgl init.d script, aside from the obvious template variables it's fairly standalone: https://github.com/crawl/dgamelaunch-config/blob/master/utils/webtiles 09:12:54 in normal circumstances if a crawl process doesn't exit within 10s something weird is going on (especially since the normal case is stale process on user exit) 09:24:28 so after reviewing all this, if I were setting this up from scratch I think my preference would be to have for the crawl process SIGQUIT = crash shutdown, SIGTERM = graceful shutdown, SIGHUP = whatever. So move current SIGTERM role to SIGQUIT, which isn't so big a change 09:25:02 it'd be yet another thing that would need a webtiles server restart 09:25:53 that looks reasonable to me, although HUP might need to do a graceful shutdown if dgamelaunch isn't configurable in that regard 09:26:04 yeah, HUP shutting down gracefully seems fine 09:26:47 re servers, i've been looking into moving to gcp or aws, specifically with some form of shared administration / billing 09:27:24 unfortunately (and unsurprisingly) this seems to be a use-case that big cloud providers do not have much support for 09:31:30 I haven't looked at gcp but when I've specced out a crawl server on aws it's way too expensive 09:32:19 and yeah, I think part of the problem is that it's just not a typical cloud use case 09:34:18 something like digitalocean is a much better bet imo, that's where I would put this hypothetical discourse reimagining of tavern 09:34:54 also where I've concerned running some kind of crawl server 09:36:16 it's probably not impossible to reimagine dcss online as a modern cloud app? I'm not sure I think this is a good idea though 09:37:12 for reference, CKO's costs for April are $34.00, presumably USD, which is indeed a little pricey 09:37:27 which is on DO 09:37:42 well, that's way cheaper than it probably would be on aws 09:39:06 how did you find that out? is there a patreon somewhere? 09:39:11 i guess it depends what you mean by a 'modern cloud app' 09:39:15 https://crawl.kelbi.org/donate/ 09:39:38 cbro is about $50 a month 09:42:11 i imagine most of this is in storage costs? 09:42:11 no idea about cao except that to connect 1TB of storage to a typical modern server would be absurd 09:42:27 I'm not sure cbro's disk is that big, that's part of the issue 09:42:30 ^status 09:42:30 14 Crawlers. CBRO disk usage=88% (135GB) | RAM usage=32% (4GB)| uptime/CPU= 09:42:30 up 137 days, 7:14, 1 user, load average: 0.42, 0.42, 0.35 (4 Cores) http://status.berotato.org 09:45:07 like, the cheapest DO droplet that has anything approaching CAO's storage is $320/mo. Though of course that's ssd...in DO you could connect enough object storage to match CAO's mostly static ttyrec archive for $20 / month 09:45:30 (which is kind of overpriced for object storage IMO, you're paying for convenience) 09:46:06 looks like it's implemented using S3 so they're probably more or less passing on S3 prices (which are high) 09:47:39 to clarify, it's disk storage that's $320/mo, and object storage that's much cheaper? 09:48:01 that's a droplet with 1T of connected SSD (it's big iin various other ways too) 09:48:42 I guess one general thought pattern I go round on any time I start to think through this stuff is: we are a small team and I personally have limited and varying time, and stuff like this can accept effectively unlimited time/effort from me as far as I can tell 09:49:25 so modernizing the server infrastructure is great and all but I'm mixed on whether it's net good for dcss overall if more and more devteam time goes into these things, with devteam size staying the same 09:51:19 so ssd is indeed exorbitantly expensive, but I'm not sure that it's necessary for most data 09:51:25 it shouldn't be 09:51:53 I was mentioning it because in some ways it's the closest modern option to how CAO is set up (though of course CAO presumably has spinning disks only) 09:52:30 i do agree that this stuff tends to be quite draining, and a lot of the pie-in-the-sky stuff is both unlikely in the near future and would not be hugely beneficial 09:53:03 but for any server in a modern configuration there's the problem of what to do with all the ttyrecs 09:53:56 (tbc CAO's size isn't a solution to that, it just postponed it by many years) 09:53:58 what I am interested is in allowing current devs to maintain the full set of crawl servers, and hopefully make them better understood 09:54:09 remove the risk of server restarts, etc. 09:54:59 the ttyrecs are indeed a bit of a problem :/ 09:55:05 maybe, but part of my point is that I don't know if I entirely agree that centralization is a net good, because right now the volunteer labor is distributed throughout a set of people that is much larger than the current active devteam 09:57:34 centralization may not be, but I think consistency would be 09:59:08 then again, centralization hopefully leads to some degree of shared access, which is also a good thing because it lowers dependencies on individual team members 09:59:48 re ttyrecs, I guess one solution is simply to not store them forever 10:00:40 I have a plan to move CAO's into my ttyrec account 10:01:14 consistency would be good but I am concerned that centralizing more stuff will actually lead to too many eggs in our basket and actual bad consequences for the server infrastructure 10:01:41 right now we have an inconsistent sytems that does limp along, more or less 10:02:34 but there are 9 servers, each with mostly their own admins; a few of those are semi-absentee 10:02:40 we can't take on that work 10:02:55 without some drastic change in how things are set up 10:03:55 your ttyrec account? 10:03:55 er 10:03:55 sorry, my b2 account 10:03:59 aah 10:04:10 where I'm already storing what we have of the cszo morgue archive 10:04:52 I do kinda think that we maybe should stop storing all ttyrecs for all time, but I'm not sure everyone agrees and I know there would be community members who are sad about that 10:05:02 *who would be sad 10:05:40 one idea might be to have a rolling period where community members could nominate ttyrecs for saving 10:06:00 and/or have time to download their own if they really want them 10:07:16 another approach would be to remove ttyrecs that are never downloaded after some grace period 10:07:32 I imagine that most ttyrecs would fall into that category 10:08:14 yeah, probably 10:10:16 i do agree that the webtiles servers do present a huge amount of work 10:10:51 that's due to the huge variety in how they're set up, and something that consistency would help with, but definitely the case now 10:12:51 I think variety is one factor, but my experience with CAO is not that that has really been a major part of what has eaten my time 10:13:06 (of course, I'm not sure how one would reunify with heavily modded servers like cwz) 10:15:10 the biggest time sink for me is simply that the whole setup is so highly bespoke/idiosyncratic in conception, that if you don't already know how to do most admin tasks, it takes a long time to figure out how to do them 10:15:33 CAO's setup is somewhat non-standard but it's not that far off vanilla dgamelaunch-config 10:16:14 that is, of course, what I'd like to fix :) 10:16:38 I don't think that is a technical problem 10:16:52 other factor for CAO in particular is just its age, there're layers of old stuff all around 10:18:11 i don't think I agree; I think there are major design flaws with the way webtiles is set up that makes maintenance more toilsome than necessary 10:18:25 which are? 10:18:29 this is only speaking from experience with VM setups, of course 10:19:00 I mean, there are many things that could be improved but a setup like this is never not going to be bespoke 10:19:21 there can't have been more than 20 crawl servers ever 10:20:14 just one example, the need to manually create folders, and the fact that no part of this setup will detect that issue and fix it automatically (or report a clear error message if auto-fixes are not possible) 10:20:27 well, to be clear I think one challenge is that much of our server infrastructure has been suffering from benign neglect for a long time because it mostly worked, that is a technical problem that can be fixed 10:20:57 other major issues are the lack of a clear / consistent base image/os, and no documentation whatsoever on installed packages / libraries 10:21:35 that is definitely true 10:22:09 that's why I started the tornado branch in the first place 10:22:50 right; now we have pinned versions, although we have yet to use them 10:22:58 re folders, the version-specific folder code you added to dgl-config is definitely good 10:23:00 so it's moving in the right direction 10:23:11 ??disjunction 10:23:11 disjunction[1/2]: Level 8 translocation spell; for a few turns, blinks or teleports stuff that's near you to be less near you 10:23:30 what if we reworked disjunction to blink things away from you when you get hit in melee? 10:23:43 it'd be less chaos roulette 11:08:39 That seems considerably weaker 11:10:52 hey all 11:11:10 (this is brianlheim from github) 11:11:22 i wanted to chat a bit about https://github.com/crawl/crawl/pull/1059 since the conversation there is already getting quite long 11:12:04 hi 11:14:06 brian2357: thanks for the work on that; if you have questions just go ahead and ask, there's usually a few people around about now but we come and go 11:15:07 ok! well, i haven't done much yet :) 11:15:52 so, i did look around to check how an MSVC build might work.. the problem is really not the crawl repo so much as the dependencies. 11:16:21 definitely, building them from source would be difficult as many have ad-hoc solution files or "special instructions in a readme"-style build procedures for MSVC 11:17:06 since both the repo and many of the dependencies are kind of unix-oriented MSYS2 seems like a better route based on what i'm seeing and also what other devs in that thread are saying 11:18:46 now, what i'm still a bit unsure of is the requirements for this PR, especially with CI.. do you want to run the same build matrix (in Travis or Github Actions) just with CMake instead of makefiles? 11:21:16 i think ideally yes 11:22:38 okay.. i was a bit confused by rawlins's comment "the plan has been to aim to merge some form of cmake building without removing make support (because of the complexity of the task we don't have any plans to convert online servers to cmake any time soon after this PR gets merged)", what does online servers refer to there then? 11:23:39 servers where you can play online, e.g. http://crawl.akrasiac.org:8080/ 11:24:45 they rebuild the latest trunk version automatically, but are difficult to make changes to 11:25:57 oh! is that a separate build pipeline than travis/gh actions? 11:27:46 i would describe it as a pipeline only in the loosest possible sense of the word... they are a bunch of jury-rigged shell scripts that we barely understand, and the server operators have very little time to maintain them 11:30:55 ok! 11:44:57 it may be possible to convert some of them to cmake eventually, but I would not take solving every build use case to be the metric for when to merge a cmake patch 11:45:17 there is no reason why cmake can't coexist with our existing build systems 11:46:14 since those largely keep chugging along with only minimal tweaking needed 11:47:14 if yall are happy to maintain both then i have no problem with that. on the project where i spend most of my time things change relatively quickly because of the dependencies we have and it's stressful sometimes just maintaining the cmake-based build system 11:47:20 yes, that's going to be the sticking point 11:47:51 maintenance that is 11:47:56 our deps don't change that often 11:48:04 if I could get it to work I would probably switch to cmake 11:48:21 btw rawlins on github = me 11:48:48 oh hi! 11:51:03 and that's good, i think it should be alright once everything is organized. Ozaq has done a really good job with this as far as i can tell 12:02:56 on the other hand, I was going to look at bundling libintl at some point 12:17:15 is that not part of libc? 12:18:11 it's needed for msys 12:19:54 this is for a hacked-up version of that i18n stuff that's been floating around on the wiki 12:20:42 so you can write format strings like "{attacker:name(THE)} {i18n.verb(v, 1)} {attacker:pronoun(POSSESSIVE)} grip on {constrictee:name(THE)}." and then swap the format strings via gettext 12:21:26 everything in {} is lua, with proxy objects for monsters, actors, items etc. that have language-specific grammar functions 12:22:45 although even without that, just having i18n support for all the plain strings would be a good start 12:54:42 03ebering02 07[miscast-simplification] * 0.25-a0-740-g8654e97: Refactor monster abjuration to be a true spell 10(2 hours ago, 2 files, 25+ 30-) 13https://github.com/crawl/crawl/commit/8654e971b0e9 12:54:42 03ebering02 07[miscast-simplification] * 0.25-a0-741-g0e34867: New Monster: nameless horror 10(50 minutes ago, 6 files, 40+ 0-) 13https://github.com/crawl/crawl/commit/0e34867035c7 12:54:42 03ebering02 07[miscast-simplification] * 0.25-a0-742-g09b25e1: Implement summoning miscasts 10(2 minutes ago, 1 file, 47+ 6-) 13https://github.com/crawl/crawl/commit/09b25e1d8be0 12:54:50 Branch pull/1341 updated to be equal with miscast-simplification: 13https://github.com/crawl/crawl/pull/1341 13:28:36 Unstable branch on crawl.akrasiac.org updated to: 0.25-a0-743-g84dc206 (34) 13:29:53 I've recently made some very small pull requests 13:31:31 I'd be willing to work on some bigger projects if one of the developers was interested 13:33:40 well, there's no end of code cleanup, bug fixes, and maintenance work to be done 13:34:53 btw just as a fyi PRs with gameplay changes might not get merged and generally take a lot longer either way 13:36:47 petercordia: I could give you a list of my favourite bug reports of mine if you like. :-) 13:46:40 I suppose I could do some bug-fixes, if they're not too complicated 13:47:51 though new features are more interesting O:3 13:52:38 aidanh, that's the reason I'd like to ask around before coding up something big. I recognise it might all be a wast of time. I discussed my 2 small pull requests in the Tavern, but the tavern doen't appear to have too much influence on actual development. 13:52:39 if you can speak a second language, internationalization is a new feature that is reasonably focused 13:55:39 it's probably a good idea to ask, yeah; I don't work much on gameplay but even if they aren't a good fit for mainline crawl, there are plenty of forks that might be willing 13:56:02 The build is still failing. (miscast-simplification - 09b25e1 #13006 : Edgar A. Bering IV): https://travis-ci.org/crawl/crawl/builds/674114242 13:56:26 petercordia: a second large-scale feature you might be interested in working on is modding support 13:56:40 although that is quite a bit larger in scope than internationalization 13:57:35 what would be the difference between a mod and a branch? 13:58:14 bcrawl & hellcrawl seem to be everything I would expect from a mod 13:59:13 basically that, but downloadable as separate files that mod mainline crawl at runtime; no recompilation necessary 14:00:15 mods would be somewhat smaller in scope; e.g. a mod for a new species, or perhaps a new school of magic, or just changing some stuff around (like your PRs) 14:01:12 there is a draft design doc at https://github.com/crawl/crawl/wiki/DCSS-Mods 14:04:08 petercordia: regarding gameplay PRs, I've got responding to yours on my todo but it takes me some time to get to them. 14:04:41 However: as a general rule of thumb, tavern is a place to get feedback on an idea or draft implementation more than it is a place to lobby for a change or conduct a poll of support 14:05:41 When making gameplay changes we look for mechanical justification over flavor justification. 14:05:57 "It would be cool" is a non-starter 14:14:27 if you want a really impactful project implement sso 14:14:44 [not actually serious, I wouldn't wish that project on anyone really] 14:15:48 petercordia I do think the smoothest way to get started is to work on bugfixes and UI improvements / fixes 14:16:37 https://github.com/crawl/crawl/wiki/User-Interface-Improvements 14:16:47 bring your shears; crawl's ui is the hairiest yak around 14:17:22 where I'd put the merfolk thing in that latter category, so is consistent with that advice 14:20:54 aidanh, the modding support project looks like it is too big for me at the moment, and probably above my skill level. 14:21:08 ebering, thank for letting me know 14:21:59 advil, I think underhound.eu already has Single sign-on, if that's what you meant 14:21:59 no worries, it's daunting for me too 14:22:09 I'll have a look at those UI issues, some of them look doable 14:22:39 One UI issue I'd like to do, but which isn't listed there, is to merge R and T 14:23:25 It's annoying to remember whether jewellery or armour is "taken off", and I know others in the tavern agree 14:25:55 single sign on means one shared account for all the servers, so it's quite a formidable task, most likely multi-year, and one we can't actually begin to work on right now anyway 16:16:19 -!- aidanh_ is now known as aidanh 17:53:55 03ebering02 07[miscast-simplification] * 0.25-a0-743-g8781be5: Update spell failure coloring 10(21 minutes ago, 2 files, 69+ 30-) 13https://github.com/crawl/crawl/commit/8781be5d8bc0 17:53:55 03ebering02 07[miscast-simplification] * 0.25-a0-744-g2672b84: fixup! Refactor and revise miscasts 10(20 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/2672b8406184 17:54:04 Branch pull/1341 updated to be equal with miscast-simplification: 13https://github.com/crawl/crawl/pull/1341 17:54:40 ebering: were you planning to go over the zot trap weighting/effect list? I can try doing so tomorrow if you're not going to get to it 17:54:53 before then, I mean 17:59:15 Killing monsters by DS Hurl Damnation doesn't count towards song of slaying bonus 13https://crawl.develz.org/mantis/view.php?id=12237 by Rogo 18:13:04 gammafunk: yes, that'll get a pass 18:23:04 Unstable branch on underhound.eu updated to: 0.25-a0-743-g84dc20635f (34) 18:42:16 aidanh: cxc is ubuntu, check the 404 page 18:51:50 The build is still failing. (miscast-simplification - 2672b84 #13008 : Edgar A. Bering IV): https://travis-ci.org/crawl/crawl/builds/674184851 19:04:41 New branch created: pull/1350 (1 commit) 13https://github.com/crawl/crawl/pull/1350 19:04:41 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1350 * 0.25-a0-738-g6f95ac3: Example webtiles systemd unit file 10(3 minutes ago, 1 file, 15+ 0-) 13https://github.com/crawl/crawl/commit/6f95ac30a42e 19:08:08 03kate-02 07* 0.25-a0-744-g404d5d3: Make WJC wall jumps only activatable via the ability 10(9 weeks ago, 9 files, 25+ 107-) 13https://github.com/crawl/crawl/commit/404d5d30d493 19:08:08 03kate-02 07* 0.25-a0-745-gac708d3: Clear water elemental engulfing when walljumping 10(27 hours ago, 1 file, 7+ 1-) 13https://github.com/crawl/crawl/commit/ac708d343770 19:08:08 03kate-02 07* 0.25-a0-746-gc9e437d: Refactor and simplify Heavenly Storm 10(6 weeks ago, 16 files, 88+ 75-) 13https://github.com/crawl/crawl/commit/c9e437d3e552 19:08:08 03kate-02 07* 0.25-a0-747-g59c31d1: Remove the pinning effect from WJC whirlwind attacks 10(2 days ago, 16 files, 14+ 48-) 13https://github.com/crawl/crawl/commit/59c31d1a9f7f 19:15:12 i didn't end up adding any other new replacement effect to whirlwind, i could see it being reasonable to mess with some other effects maybe but that can probably be thought about separately 19:25:17 ohhh why the cap :( :( 19:25:30 if people can get 50 slay clearing vaults 5 why stop them? 19:30:00 New branch created: pull/1351 (1 commit) 13https://github.com/crawl/crawl/pull/1351 19:30:00 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1351 * 0.25-a0-744-g4ba0c6e: Send DCSS processes SIGABRT rather than SIGTERM 10(17 minutes ago, 1 file, 4+ 4-) 13https://github.com/crawl/crawl/commit/4ba0c6e6bc53 19:37:08 -!- DarwinFreyson is now known as DarwinElf 19:44:43 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-747-g59c31d1a9f (34) 19:59:02 03ebering02 07[miscast-simplification] * 0.25-a0-748-ga40a96f: Adjust Trog wrath 10(7 days ago, 1 file, 19+ 13-) 13https://github.com/crawl/crawl/commit/a40a96fbb7e5 19:59:02 03ebering02 07[miscast-simplification] * 0.25-a0-749-g571808c: Adjust Lugonu wrath 10(7 days ago, 1 file, 6+ 11-) 13https://github.com/crawl/crawl/commit/571808c64f26 19:59:02 03ebering02 07[miscast-simplification] * 0.25-a0-750-gf8edb28: Adjust Cheibriados wrath 10(7 days ago, 1 file, 9+ 16-) 13https://github.com/crawl/crawl/commit/f8edb28dd6df 19:59:02 03ebering02 07[miscast-simplification] * 0.25-a0-751-g7d74140: Adjust Fedhas wrath 10(7 days ago, 3 files, 23+ 12-) 13https://github.com/crawl/crawl/commit/7d7414004a6f 19:59:02 03ebering02 07[miscast-simplification] * 0.25-a0-752-g2736e7e: Decouple distortion unwields from miscasts 10(6 days ago, 5 files, 41+ 36-) 13https://github.com/crawl/crawl/commit/2736e7e68b84 19:59:02 03ebering02 07[miscast-simplification] * 0.25-a0-753-g47bc34d: Remove miscasts from Yred wrath 10(5 days ago, 1 file, 12+ 5-) 13https://github.com/crawl/crawl/commit/47bc34da0a4b 19:59:02 03ebering02 07[miscast-simplification] * 0.25-a0-754-g86a13c9: Remove dead code from chaos melee miscasts 10(5 days ago, 5 files, 1+ 123-) 13https://github.com/crawl/crawl/commit/86a13c932734 19:59:02 03ebering02 07[miscast-simplification] * 0.25-a0-755-g336ef83: Refactor and revise Mummy death curses 10(4 days ago, 5 files, 263+ 7-) 13https://github.com/crawl/crawl/commit/336ef83941f6 19:59:03 03ebering02 07[miscast-simplification] * 0.25-a0-756-gfb0a38e: Adjust the scythe of Curses 10(4 days ago, 2 files, 7+ 24-) 13https://github.com/crawl/crawl/commit/fb0a38e7c02c 19:59:03 03ebering02 07[miscast-simplification] * 0.25-a0-757-g03a4b90: Remove miscasts from Kiku wrath 10(4 days ago, 2 files, 13+ 17-) 13https://github.com/crawl/crawl/commit/03a4b90f3d04 19:59:03 ... and 8 more commits 19:59:11 Branch pull/1341 updated to be equal with miscast-simplification: 13https://github.com/crawl/crawl/pull/1341 19:59:59 rebased and force-pushed 20:00:13 %git … 20:00:13 Could not find commit … (git returned 128) 20:00:21 %git c46e80e 20:00:21 07ebering02 * 0.25-a0-760-gc46e80e: Decouple Zot trap effects from miscasts and revise 10(33 hours ago, 1 file, 52+ 27-) 13https://github.com/crawl/crawl/commit/c46e80e7f089 20:00:41 gammafunk: this has the updated table including sgd, tentacle, and twister along side recall 20:20:55 New branch created: pull/1352 (1 commit) 13https://github.com/crawl/crawl/pull/1352 20:20:55 03riverfiend02 {GitHub} 07https://github.com/crawl/crawl/pull/1352 * 0.25-a0-748-ga4b70fa: New Hep+Dith shared overflow altar 10(2 minutes ago, 1 file, 25+ 1-) 13https://github.com/crawl/crawl/commit/a4b70fae2aba 20:37:36 03ebering02 07* 0.25-a0-748-ga40a96f: Adjust Trog wrath 10(7 days ago, 1 file, 19+ 13-) 13https://github.com/crawl/crawl/commit/a40a96fbb7e5 20:37:36 03ebering02 07* 0.25-a0-749-g571808c: Adjust Lugonu wrath 10(7 days ago, 1 file, 6+ 11-) 13https://github.com/crawl/crawl/commit/571808c64f26 20:37:36 03ebering02 07* 0.25-a0-750-gf8edb28: Adjust Cheibriados wrath 10(7 days ago, 1 file, 9+ 16-) 13https://github.com/crawl/crawl/commit/f8edb28dd6df 20:37:36 03ebering02 07* 0.25-a0-751-g7d74140: Adjust Fedhas wrath 10(7 days ago, 3 files, 23+ 12-) 13https://github.com/crawl/crawl/commit/7d7414004a6f 20:37:36 03ebering02 07* 0.25-a0-752-g2736e7e: Decouple distortion unwields from miscasts 10(6 days ago, 5 files, 41+ 36-) 13https://github.com/crawl/crawl/commit/2736e7e68b84 20:37:36 03ebering02 07* 0.25-a0-753-g47bc34d: Remove miscasts from Yred wrath 10(5 days ago, 1 file, 12+ 5-) 13https://github.com/crawl/crawl/commit/47bc34da0a4b 20:37:36 03ebering02 07* 0.25-a0-754-g86a13c9: Remove dead code from chaos melee miscasts 10(5 days ago, 5 files, 1+ 123-) 13https://github.com/crawl/crawl/commit/86a13c932734 20:37:36 03ebering02 07* 0.25-a0-755-g336ef83: Refactor and revise Mummy death curses 10(4 days ago, 5 files, 263+ 7-) 13https://github.com/crawl/crawl/commit/336ef83941f6 20:37:36 03ebering02 07* 0.25-a0-756-gfb0a38e: Adjust the scythe of Curses 10(4 days ago, 2 files, 7+ 24-) 13https://github.com/crawl/crawl/commit/fb0a38e7c02c 20:37:36 03ebering02 07* 0.25-a0-757-g03a4b90: Remove miscasts from Kiku wrath 10(4 days ago, 2 files, 13+ 17-) 13https://github.com/crawl/crawl/commit/03a4b90f3d04 20:37:36 ... and 8 more commits 21:14:10 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-765-g8a17258929 (34) 21:31:11 oh no, plutonium sword without poly other is very sad imo... 21:33:16 kate-: I wouldn't object to plut sword getting a slightly more involved effect, but I wanted miscasts to be simple 21:34:18 yeah definitely fair, overall it definitely looks like a good set of changes (and as an unrand it can probably justify having a small separate chance for poly or something) 21:43:37 03kate-02 {GitHub} 07* 0.25-a0-766-g0ecfff5: Tweak wording and fix a typo 10(11 seconds ago, 1 file, 3+ 3-) 13https://github.com/crawl/crawl/commit/0ecfff59826d 21:43:57 The formulas might need some tuning; hell effects probably need their own careful look but this set of changes was getting pretty big 21:47:47 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-766-g0ecfff5982 (34) 21:49:16 03riverfiend02 {GitHub} 07https://github.com/crawl/crawl/pull/1352 * 0.25-a0-749-g7e6c16b: Update overflow.des 10(3 minutes ago, 1 file, 0+ 1-) 13https://github.com/crawl/crawl/commit/7e6c16b03ff0 21:51:42 The build is still failing. (miscast-simplification - 8a17258 #13013 : Edgar A. Bering IV): https://travis-ci.org/crawl/crawl/builds/674210683 22:27:30 aidanh fyi MITRE has issued CVE-2020-11722 for the security bug. When you finalize your security advisory, let Github know so they don't issue a duplicate CVE. 23:29:26 The build was broken. (master - 0ecfff5 #13017 : Kate): https://travis-ci.org/crawl/crawl/builds/674229666