03:31:34 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-5208-geafff8c3b6 05:07:40 Unstable branch on crawl.akrasiac.org updated to: 0.32-a0-1210-g732af37 (34) 06:13:09 03dolorous02 07* 0.32-a0-1211-g5cb19619eb: Add more minor alchemy miscast messages. 10(19 minutes ago, 1 file, 2+ 0-) 13https://github.com/crawl/crawl/commit/5cb19619eb05 06:43:57 I have a few questions regarding setting up a WebTiles server 06:44:04 03dolorous02 07* 0.32-a0-1212-g69e7503642: Don't use mprf() when mpr() will do. 10(2 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/69e7503642da 06:44:12 1. Most public servers host game data using Apache. Can sequell handle an nginx-style directory index page? 06:44:30 2. Regarding this testing container project: https://github.com/crawl/dgamelaunch-config/tree/master/utils/testing-container, is it acceptable to use this setup as the base configuration for a production server? Is it true that in a DGL session, users cannot execute arbitrary code? Also, is it safe from container escaping issues? I have read the 06:44:30 code and it doesn't seem to have major issues, but there are comments suggesting it might not be suitable for production, and that the crawl-dev account might have too many permissions for production use. 06:44:39 3. Does the ttyrec recording feature significantly impact the real-time performance of the server? Is it necessary to use SSDs for this due to potential performance issues? 06:48:23 <09gammafunk> ASCIIPhilia: Yes, you can use nginx. I use it to host the official server CDI. Sequell can handle dir listings from non apache servers 06:49:51 ttyrecs record at human speeds, it's a tiny fraction of what spinning rust can do 06:52:53 <09gammafunk> As for the container, advil can comment on running that in production, but you can fully install the dgl setup without going through the container. That is what all current DGL servers do. And for crawl-dev, I recommend disabling its account password so that logins can only happen via eg sudo or possibly an ssh key if you do want want remote access to that account 06:54:47 Pinkbeast I heard that the administrator of the CWZ server introduced SSDs to improve I/O performance (probably due to ttyrec I/O from multiple sessions). I also heard that CWZ transfers data files from SSD to HDD every day at 5 AM KST. 06:54:47 However, I have been wondering if ttyrec I/O really demands SSDs due to heavy load. 06:54:48 (I'm also a little skeptical about this part.) So, I wanted to ask for opinions from anyone who manages a production server. 06:55:13 <09gammafunk> I think the comments you're referring to are talking about how crawl-dev can escalate privileges via a sudo entry it recommends, and that is true, but don't give people access to the crawl-dev account unless you want them to also have root 06:57:00 <09gammafunk> For ttyrecs, it is a lot of disk activity as the number of users grows, and CAO has had issues with this and had to disable ttyrecs. It does not have an SSD, for example 06:57:41 @gammafunk It's great to hear that nginx can also be used. 06:57:42 So, does that mean that other servers, except for the one operated by advil, are running dgl installed directly on the host rather than using containers? 06:57:49 <09gammafunk> But cao has older disks in general 06:59:00 <09gammafunk> All dgl servers including CAO (the one advil primarily operates) use dgl directly with no container 06:59:23 <09gammafunk> We have no official server using the container as far as I know 07:00:35 <09gammafunk> I mentioned advil because advil made that container and understands it better than I do 07:04:16 I tried to set up the server using doku wiki documentation on Ubuntu 18.04.6, but too many dependency issues troubled me. Therefore, if it's safe, I would like to build a production server based on the testing container (as it seems easier to manage in the future). 07:05:41 Rather than asking for a thorough technical verification, I'm just wondering if it's okay to use it. 07:13:06 If the setup goes well, I plan to post a one-stop docker-compose setting on GitHub that configures the web server and multiple game builds simultaneously. 07:40:49 Yes, that's right. From what I understood about the whole process, users will only use the account where dgamelaunch is the shell, and if I set the permissions to prevent this account from doing anything dangerous, there shouldn't be any problems. However, since this kind of setup is unfamiliar to me, I wanted to ask if there are any exceptions to 07:40:50 be aware of. 07:42:53 @gammafunk I couldn't read the message above. Sorry. haha 07:51:45 <05icoson> the testing container is really not designed for production 07:51:57 <05icoson> I believe one non-official server does have a production container setup of some kind 07:53:34 <05icoson> https://github.com/Rytisgit/dgamelaunch-dcss-forks-server (username in discord at least is Sentei) 07:53:50 <05icoson> iirc it's based on the testing container 07:54:35 <05icoson> (btw I think I may show up as a different username in irc but this is advil speaking) 07:55:34 <05icoson> the testing container is really designed to exactly replicate a standard (chroot / no-container) setup but inside a docker container, so that it can be tested in ci 07:56:06 <05icoson> which makes it quite complex in ways that are unnecessary, because of what's needed to set up a chroot in a docker container 08:02:21 You're right, it's unfortunate but understandable that privileged permissions are required. Still, the configuration is indeed complex even in a non-container environment, so I think it's a very impressive and attractive setup. 08:04:30 *Things Sacrificed to set up chroot & dgl 08:06:09 03dolorous02 07* 0.32-a0-1213-gdb9be81004: Remove unneeded blank line. 10(18 minutes ago, 1 file, 0+ 1-) 13https://github.com/crawl/crawl/commit/db9be81004a3 08:06:09 03dolorous02 07* 0.32-a0-1214-g6251899524: Fix spelling. 10(15 minutes ago, 9 files, 14+ 14-) 13https://github.com/crawl/crawl/commit/6251899524a6 08:06:09 03dolorous02 07* 0.32-a0-1215-g540c3cb408: Fix obsolete orb ego: channelling -> energy. 10(4 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/540c3cb4082a 08:07:03 What do you think are the points that make it unsuitable for production use? I don't think the performance overhead due to containerization is significant, and I believe additional configurations such as miscellaneous settings, game build setups, and web server configurations can be added without issues. Do you have any concerns? 08:12:05 I'm considering a production scale similar to the CWZ server, with around 1-5 users during regular times and 10-15 users at peak times. 08:21:05 03dolorous02 07* 0.32-a0-1216-gab2d129dd7: Add a message for when the player finds all gems. 10(2 hours ago, 1 file, 30+ 0-) 13https://github.com/crawl/crawl/commit/ab2d129dd713 08:24:10 04Build failed for 08master @ ab2d129d 06https://github.com/crawl/crawl/actions/runs/9148678068 08:26:25 03dolorous02 07* 0.32-a0-1217-gbb3f4b5168: Checkwhite 10(2 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/bb3f4b51682f 08:36:14 <06dolorous_84348> Now, I suppose, some super-hardcore players will see how many of the 360 miscellaneous colors they can make all the gems sparkle. 08:56:37 <09gammafunk> ASCIIPhilia: You may want to hear from @sentei , who is running a containerized server as advil mentioned above. Sentie is in this discord and can answer questions about any caveats for running webtiles via a container. It looks like sentei's container even supports the dgamelaunch-config scripts we use for our officials servers, and it presumably supports normal dgamelaunch/ssh connections as well. For running a server, it looks like 08:56:37 a better starting point than the testing container: https://github.com/Rytisgit/dgamelaunch-dcss-forks-server 08:57:34 <09gammafunk> It looks like it's also using a chroot 08:58:01 <09gammafunk> so presumably it has the same container issues associated with running a chroot within the container as our testing container would 08:58:25 <06pleasingfungus> i wonder how much work it'd be to upgrade to c++23 08:58:35 <06pleasingfungus> i'd really really really like https://en.cppreference.com/w/cpp/io/print 08:58:52 <09gammafunk> The main issue has always been that we have several official servers that run older OSes 08:59:19 <09gammafunk> e.g. CAO and perhaps CUE 08:59:43 <06pleasingfungus> sure 09:01:27 @gammafunk Thank you for the detailed response. 09:02:05 That project definitely looks like a better starting point. I'll keep it in mind. Thanks 09:03:53 <09gammafunk> ASCIIPhilia: No problems, and thanks for doing the work to start up a new server! If you're successful, I'm sure a lot of players in the Asia-Pacific region will be happy to have a better place to player crawl 09:12:08 This isn't related to the server, but I've recently made something like this (I'm not sure if the translation will be accurate). It's a sort of WebTiles third-party plugin framework (a successor to the CWZ add-on script). 09:12:08 https://papago.naver.net/website?locale=ko&source=ko&target=en&url=https%3A%2F%2Fgall.dcinside.com%2Fboard%2Fview%2F%3Fid%3Drlike%26no%3D459000 09:12:09 https://papago.naver.net/website?locale=ko&source=ko&target=en&url=https%3A%2F%2Fgall.dcinside.com%2Fboard%2Fview%2F%3Fid%3Drlike%26no%3D459204 09:12:09 Once it's more developed, I plan to promote the project with proper documentation, including an English version. 09:13:35 <03wheals> since crawl doesn't print to stdout or files very often, i assume std::format from c++20 would be enough 09:13:42 <03wheals> not that we support that either 09:14:49 <06pleasingfungus> seems plausible 🙂 09:15:39 <06pleasingfungus> maybe i should talk to rax and see how much work it'd be to get a compiler that'd support newer cpp versions 09:15:44 <06pleasingfungus> i don't know how to talk to rax, tho. 09:18:02 <06pleasingfungus> asciiphilia: btw, ubuntu 18 is past end of life, iirc, so i wouldn’t recommend using it for a new host. 09:18:51 <06pleasingfungus> ideally you’d use ubuntu 24 09:20:23 Yes, I usually use the latest Ubuntu LTS build. However, it seems that the installation guide was written when that version was in use, so I temporarily used it to test the non-container environment setup (to get familiar with server settings). 09:20:57 Even though I used 18.04.6, I still encountered a lot of dependency errors... 09:21:02 <06pleasingfungus> gotcha 09:21:06 <06pleasingfungus> that makes sense 09:27:58 <06pleasingfungus> but hopefully you’d switch to 24 for the server once you start it 09:28:18 <06pleasingfungus> (thank you very much for volunteering to run a server, btw!) 09:34:27 I'm currently running the https://webtiles.nethack.live server in Gyeonggi-do, South Korea, so the DCSS WebTiles server will also be operated 24/7 on the same server. However, the server setup seems more complex than expected, and I'm considering setting up a backup service using AWS S3 Glacier for data stability. The actual server setup might be 09:34:27 completed late in the second half of this year. It could be sooner or later, but it's uncertain. (Sadly, I have my graduate school entrance exams this year.) 09:36:44 Anyway, thank you all so much for the detailed answers to my questions. 09:44:09 New branch created: pull/3824 (1 commit) 13https://github.com/crawl/crawl/pull/3824 09:44:09 03NormalPerson702 07https://github.com/crawl/crawl/pull/3824 * 0.32-a0-1199-g641605e8df: Fix Hellfire Mortar power cap (#3820) 10(6 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/641605e8df98 09:49:30 03NormalPerson702 07https://github.com/crawl/crawl/pull/3824 * 0.32-a0-1199-gfe1483a181: Fix Hellfire Mortar power cap (#3820) 10(12 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/fe1483a1818e 10:55:48 <02sentei> It actually has the chroot removed, since advil had left a coment that if it was going to be containerised it wouldnt be necesary anymore. 10:57:38 <02sentei> @asciiphilia if you have questions regarding it, I would be happy to help. 10:58:11 <09gammafunk> I do not recommend using an older version of ubuntu for dgamelaunch, yes. Everything should work perfectly fine with a current debian (or ubuntu or other debian-derived distro) OS 10:58:21 <09gammafunk> ASCIIPhilia 10:58:22 <09gammafunk> ^ 10:58:23 <09gammafunk> oops 10:59:16 <09gammafunk> That guide was simply written a long time ago and updated a bit over the years, but you shouldn't need anything old; CDI is running ubuntu 22.04, for example, which is nearly the current LTS release (24.04 was released only very recently) 11:00:39 <09gammafunk> That's good. Have you updated the dgamelaunch-config scripts to work accordingly? I guess you're still using the templating functionality of those scripts, but it's just publishing to a filesystem location that isn't in any kind of chroot? 11:01:53 03DracoOmega02 07* 0.32-a0-1218-g91856e71ab: Buff / Fix Rimeblight 10(55 seconds ago, 7 files, 60+ 36-) 13https://github.com/crawl/crawl/commit/91856e71ab78 11:03:43 <09gammafunk> What really needs to happen on CAO is a dist upgrade, something advil asked rax about quite a while ago, but I'm not sure if he received a response. It's non-trivial from a sysadmin point of view, because CAO isn't virtual, and if something happens during a dist upgrade that breaks the controlling ssh connection, the server will potentially be inaccessible unless someone can login e.g. with a local terminal on the machine in the data 11:03:44 center. And I'm not sure at all about the state of CAO's hardware in terms of a dist upgrade being practical; advil might know more about that 11:06:22 <09gammafunk> ASCIIPhilia: so to summarize regarding a recent OS, please try to use a reasonably recent ubuntu/debian distro and if you encounter any dependency problems, just ask about them in this channel. I don't think anyone has tried ubuntu 24.04 since that was only released in late April, but in any case we can help you troubleshoot dependency problems etc. It can just take a while for people here to respond 11:08:17 <06pleasingfungus> gotcha. i’ll wait on advil before investigating further 11:10:38 <09gammafunk> although, now that I say this, apparently ebering did this a while ago, I'm just not sure on which servers it was done: https://github.com/crawl/crawl/wiki/CPP-17-Upgrade 11:11:35 <09gammafunk> looks like it installed ubuntu 19.10 into the chroot? Which I'm not sure how that plays with the dgamelaunch-config build process, since usually if there's a version mismatch between the out-of-chroot build directory and the chroot, problems arise 11:14:00 <09gammafunk> oh I see, this script actually changes the crawl-dev account dgamelaunch setup to build inside the new chroot 11:16:57 <06pleasingfungus> ooh 11:17:00 <06pleasingfungus> i’m optimistic 11:18:31 <09gammafunk> wasn't used on CAO, and I'm not sure it's a wise procedure, but I appreciate its ambition. I'd definitely leave the GCC upgrade on CAO project to advil, at least in terms of a recommendation of how to do it if a dist upgrade isn't possible 11:18:45 <09gammafunk> but we're apparently at 4.7.2 on CAO 11:22:48 <02sentei> The scripts do work, I pass in docker container root as the chroot. Regarding publishing I did set up 4 volumes for the player and game data which is supposed to be persistent in the real filesystem, but I'm not sure if those included the files updated using dgl publish, I might have thought that its better to track those kind of changes in a git repo and just pull in a new docker image with the updates of the scripts and restart. 11:24:19 <02sentei> At least thats what I have been doing so far. 11:29:44 <06pleasingfungus> looks like ubuntu 24 would support the stuff i want. 22, maybe 12:13:31 @sentei Is it correct to understand this as a version of the container that has removed the use of chroot? It seems that some parameters have changed compared to https://github.com/crawl/dgamelaunch-config/blob/master/utils/testing-container/Dockerfile (but dgamelaunch-config feels a bit more up-to-date, since it's using the 22.04 image). 12:14:20 @gammafunk Okay, next time I set up a non-container environment, I'll try using version 24.04. 12:15:02 https://github.com/Rytisgit/dgamelaunch-dcss-forks-server/blob/master/utils/testing-container/Dockerfile v.s. https://github.com/crawl/dgamelaunch-config/blob/master/utils/testing-container/Dockerfile 12:17:23 Oh, the docker-compose configuration is in the same directory. That's interesting and seems like it will be very helpful for reference. 12:38:27 <02sentei> @asciiphilia, yes my fork has the chroot removed from the container. I was the only one hosting this version, so I didnt feel the need to bump the version. 12:45:04 @sentei Is the crawl-forks repository a collection of Crawl versions that are no longer being updated? Some Crawl fork projects seem to pull directly from upstream, but why don't the forks in crawl-forks do the same? 12:50:44 <02sentei> @asciiphilia yes the forks in the crawl-forks repo are those which are old and missing a critical security patch, so I made the repo to gather them and apply the patch in a single location 12:52:40 @sentei Oh, it's because of the security patch. It must have been a pretty critical issue if you had to gather them separately to apply the patch. 12:56:50 @sentei I was trying to do all the work myself similarly to what you did, but the parts I needed are so well organized that I'm very happy. The work might progress more easily than I thought. 12:58:01 <02sentei> Great to hear, looking forward to a dockerised official server 13:05:23 Gandelf (L1 SpCj) ASSERT(valid_cursor_pos(pos.x, pos.y, region)) in 'libutil.cc' at line 404 failed. (invalid cursor position -34,9 in region 2, should be 3,9 in region 3) (D:1) 13:08:12 Oh, and I've been curious about something for a while. Is it possible to perform a seamless trunk version update (where the server doesn't go down for a while or users don't have to quit their games)? Some servers stop the server to update the trunk (like CWZ), while other servers seem to update the trunk almost seamlessly. I'm curious if updating 13:08:13 the trunk without downtime causes any issues for the server. 13:08:13 I don't know the structure of DCSS Webtiles well, but even if a new game build binary overwrites the existing trunk binary location, can users who are already in a trunk game continue to play without issues because the binary is fully replicated in memory? 13:10:08 yes. that's an OS thing 13:10:20 it's replicated either in memory or in swap 13:11:52 cool 13:12:07 <09gammafunk> Yes, with a dgamelaunch-config setup that's properly configured, you can have many versions of trunk installed. What happens for any given trunk user is that, as long as their save is present (i.e. the game hasn't ended, so the save file still exists), that version of trunk is guaranteed to still have its directory. Once there are no more users for a given version of trunk, it's eligible for deletion according to a script that can be 13:12:07 run periodically (e.g. via cron) that deletes unused trunk versions. And when a user resumes their game, the dgamelaunch-config game launcher sees if there's a newer version of trunk installed, and if so it gives them a Y/N prompt to upgrade their game to the more current version. If the player accepts, the game now uses the new trunk version, otherwise they remain on their current version 13:13:18 <09gammafunk> You can configure things to "force" transfers to the latest version always, although I think this can run into issues when trunk is upgraded. The CPO server worked this way, but it didn't use dgamelaunch-config 13:14:57 @gammafunk Ah, I thought the built binaries were continually overwritten in one place. Are they actually generated in new directories based on something like commit hashes? 13:15:08 <09gammafunk> That's right 13:15:21 <09gammafunk> each unique commit/hash that gets installed gets its own directory 13:15:34 actually I think that depends on server config? kelbi used to do forced upgrades iirc 13:15:42 <09gammafunk> which is only deleted when no more users have saves for that version (this happens via the cron script) 13:16:00 <09gammafunk> yeah, I just forget if that's configurable in dgamelaunch-config 13:16:26 (with rebuilds every 15 minutes it'd have had far too many versions in flight otherwise) 13:16:30 <09gammafunk> but I think you'll always have a potential issue with in-progress games that are currently running if you have any setup (dgamelaunch-config or otherwise) only ever have one trunk installed 13:16:57 <09gammafunk> well CDI is running with 15 minute rebuilds and does not force transfer 13:17:19 Morgoth2127 (L7 TrMo) Crash caused by signal #6: Aborted (D (Sprint)) 13:17:43 @gammafunk Yes, I think that method seems much smarter than just overwriting everything indiscriminately. 13:18:20 <09gammafunk> You do have to be careful to periodically delete those old trunk versions with the dgamelaunch-config scripts if you allow multiple trunk versions 13:18:31 <09gammafunk> otherwise they slowly build up and are unused etc 13:19:29 <09gammafunk> I wonder if e.g. kelbi just allowed multiple installs so long as there were active saves and then always force transfered instead of having the dialog 13:20:00 <09gammafunk> so then it wouldn't run into the issue of messing up games that were active at the time of trunk rebuild/upgrade 13:20:00 Rebuilding every 15 minutes without forcing transfers is the most user-friendly approach, though it might use a bit more resources. 13:20:18 <09gammafunk> right, and you don't have to rebuild every 15 minutes regardless, it should be said 13:20:23 <09gammafunk> rebuild time is totally up to you 13:20:39 <09gammafunk> for CWZ, I think hong had local patches he liked to apply, so automatic rebuild wasn't even possible? 13:21:14 <09gammafunk> But if you're running unmodified trunk and don't have local patches that may fail to merge, you should be fine using the standard dgamelaunch-config scripts for seamless rebuilds 13:22:11 Actually, I don't know much about the structure of Webtiles, but I thought that using something like webhooks would be the most efficient approach. 13:22:46 <09gammafunk> well, dgamelaunch-config stuff that runs periodically is typically run via crontab 13:22:56 <09gammafunk> the crontab of the crawl-dev user, specifically 13:23:09 <09gammafunk> of course, you can set up some other scheme 13:24:09 In the case of hong, it seems more like the server has been running for a long time, so exceptional situations keep happening, and there wasn't much choice rather than him just liking to apply local patches. 13:26:08 I recently got in touch with hong, and I heard that many environmental issues are causing a lot of headaches because the server is so old. Formatting and setting it up anew seems difficult because the settings and configurations are too intricately tangled. 13:27:24 I heard that CWZ is currently experiencing intermittent delays (500-800ms frequently, 2-4s less frequently) and they are troubleshooting this issue. 13:28:50 After discussing together, it doesn't seem to be a network issue. When the delay occurs, there is an unusual increase in the CPU usage of the Tornado web server. We couldn't figure out why this problem is happening. 13:30:05 why does this remind me of CAO? 13:30:43 Did CAO have a similar issue? 13:30:53 does, ongoing 13:31:05 seems more related to disk in CAO's case, though 13:35:09 Sometimes, I feel like servers are like ticking time bombs waiting to explode in the future 13:35:53 When first introduced, they work great and look impressive, but as they age, they start creaking and eventually go boom! 13:36:54 <06pleasingfungus> such is technology 🙂 13:57:39 03nicolae02 07* 0.32-a0-1219-g78dc6a7b8a: Fix up some shop vaults and add some more 10(4 minutes ago, 1 file, 92+ 36-) 13https://github.com/crawl/crawl/commit/78dc6a7b8a02 14:39:01 <06regret-⸸nde※> * vaguely gives a thumbs-up and then returns her attention back towards writing up a changelog * 14:57:17 <06mumra> Considering moving the miscast strings into their own txt file 14:59:10 <06mumra> The way the randomised colours are currently being looked up there, it's doing numerous calls to the txt database every time there is a miscast, and in the majority of cases the colour is not even being used. At least if they were all already in txt they are only evaluated if needed 14:59:25 <08nicolae> i've got a bunch of crypt decor vaults lying around from earlier, i might just dump them into trunk instead of making a branch since there's nothing to really balance 15:00:53 <06mumra> It is maybe not obvious but getSpeakString is doing a huge amount of work -- in the case of colour lookups it is taking the entire text entry (with all 100+ colours), splitting it on \n and pushing all the colours into a vector (and the weights into a 2nd vector) 15:01:29 <06mumra> Then performing a random weighted selection on the vector 15:09:08 <06mumra> Which seems quite inefficient for a string that's probably being thrown away anyway (it is at least probably better to check for the existence of @any_colour@ etc before actually invoking getSpeakString ... alternately i suppose we could publicly expose _call_recursive_replacement in database.cc as this can achieve the same thing, checking for @ tokens and then looking up the corresponding text) 15:12:10 <06pleasingfungus> moving to another db seems fine, but i’m really skeptical of performance optimizations made without profiling 15:15:07 <06pleasingfungus> like i see no particular reason to expect the operations you described are very expensive 15:34:51 <06mumra> Sure and it's not like this is something frame critical anyway or called super frequently ... but at least in my understanding of string operations and vectors, given colourname.txt is 5kb (which is mostly the misc colours) it's going to be allocating 5kb for the split vector, then a further 5kb to copy the strings into another vector, each time it hits the @misc_colour@ lookup. I know 10kb is not huge in the scheme of things but it 15:34:51 might be slightly surprising that this lookup costs that much 15:38:36 Unstable branch on underhound.eu updated to: 0.32-a0-1219-g78dc6a7b8a (34) 15:59:56 <05icoson> re cao, rax was willing in principle but it's really not an easy project, and so I think has never really found the time to do it. I do ping every once in a while to see if it's possible 16:01:03 <05icoson> I was able to get that chroot setup working on cao but it ended up being to rube goldberg for me to really want to deal with, and also, it's kinda papering over all the non-crawl reasons why cao really should get some sort of update 16:01:12 <05icoson> so I'm not enthused about that solution 16:04:18 <05icoson> ASCIIPhilia: re webtiles cpu usage, there are a number of fixes to how idle connections are timed out from Mar or so that made a huge difference, if those aren't in cwz they really should be 16:07:54 they left several hours ago 16:09:41 <06dolorous_84348> For the record, I was just modeling the code for it after what was already there. Player hands aren't in every message, either, and yet the substitutions for those are always done too. 16:10:32 <06dolorous_84348> And those aren't a database lookup, but still. 16:25:43 <06dolorous_84348> They could be at some point, given the datification of player stuff. 16:40:48 03nicolae02 07* 0.32-a0-1220-gf720f7fe4a: Decor vaults for Crypt 10(2 minutes ago, 1 file, 434+ 0-) 13https://github.com/crawl/crawl/commit/f720f7fe4af3 16:52:33 03regret-index02 07* 0.32-a0-1221-g9412f0fc87: Changelog for 0.32 (up to 78dc6a7) 10(9 minutes ago, 1 file, 221+ 0-) 13https://github.com/crawl/crawl/commit/9412f0fc870e 16:52:33 03regret-index02 07* 0.32-a0-1222-gd1fd03315b: Nudge around some listings and documentation 10(80 seconds ago, 2 files, 17+ 17-) 13https://github.com/crawl/crawl/commit/d1fd03315b74 16:53:25 <06regret-⸸nde※> ...still need to work on Crypt:3 Shop so the changelog can mention Crypt's got more loot that previously. 17:00:18 @icoson Thanks for sharing the information. I'll talk to hong about it. 17:17:05 03regret-index02 07* 0.32-a0-1223-g8efb30d515: Fiddle with the changelog's future (various) 10(54 seconds ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/8efb30d5153c 17:22:55 <09gammafunk> r-i, just fyi we usually put something like "through 0.32-a0-1223-g8efb30d515" in those parenthesis at the top for trunk versions, mostly so in-games readers of the changelog can compare to their current/local version 17:25:37 03dolorous02 07* 0.32-a0-1224-g2cd1f3e945: Fix typos and spelling. 10(2 minutes ago, 1 file, 9+ 9-) 13https://github.com/crawl/crawl/commit/2cd1f3e94598 17:25:37 03dolorous02 07* 0.32-a0-1225-g48eddae406: Clarify new Ijyb's description. 10(64 seconds ago, 1 file, 2+ 1-) 13https://github.com/crawl/crawl/commit/48eddae40661 17:43:44 03dolorous02 07* 0.32-a0-1226-gd8addd5893: Fix description of DS foul flame mutation. 10(2 minutes ago, 1 file, 3+ 2-) 13https://github.com/crawl/crawl/commit/d8addd589319 17:46:57 <06dolorous_84348> One thing the changelog updates made me double check: apparently the proper Australian spelling is "bedevilled", not "bedeviled"? 17:49:55 <09gammafunk> beasmodeused 17:54:38 <08nicolae> @regret-⸸nde※ what other areas did you still want decors for? i have a note asking for "~14 shoals or snake or orc decorative vaults") 17:55:07 <08nicolae> bazaar project was on hold due to clinical brain death but i might have it in me to get going on that this week, no promises 17:58:40 03dolorous02 07* 0.32-a0-1227-g2ebb5e6a31: Update monster bot for AF_SWARM. 10(6 minutes ago, 1 file, 3+ 0-) 13https://github.com/crawl/crawl/commit/2ebb5e6a31e4 17:58:40 03dolorous02 07* 0.32-a0-1228-gf56e53c086: Colourise drag and swarm in monster bot. 10(5 minutes ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/f56e53c0865a 18:02:31 <06dolorous_84348> Incidentally, is there some way to add a monster bot build to the ci checks? (Ge0ff brought it up a while back too.) 18:05:05 03dolorous02 07* 0.32-a0-1229-g1eea0c6e6f: Fix spelling: bedeviled -> bedevilled. 10(30 minutes ago, 33 files, 106+ 106-) 13https://github.com/crawl/crawl/commit/1eea0c6e6fa6 18:09:22 03dolorous02 07* 0.32-a0-1230-g14f4ba10ed: Add 0.32 up-through version to changelog. 10(2 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/14f4ba10ed8e 18:09:58 <06pleasingfungus> @regret-⸸nde※ thanks for doing the changelog! 🙂 18:50:58 03dolorous02 07* 0.32-a0-1231-g3a8b7b03c6: Add sometimes-pasta-adjacent food to artefacts. 10(2 minutes ago, 1 file, 3+ 0-) 13https://github.com/crawl/crawl/commit/3a8b7b03c6a8 19:54:15 03dolorous02 07* 0.32-a0-1232-gb12068d400: Add correct 0.32 up-through version (oops). 10(74 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/b12068d4006c 20:33:55 03dolorous02 07* 0.32-a0-1233-gb66479f5cc: Mention Fedhas' protecting friendly plants. 10(2 minutes ago, 1 file, 2+ 1-) 13https://github.com/crawl/crawl/commit/b66479f5cc7a 22:33:39 03dolorous02 07* 0.32-a0-1234-gac1fde5492: Fix typos, spelling, and wording. 10(86 seconds ago, 1 file, 6+ 5-) 13https://github.com/crawl/crawl/commit/ac1fde549253 22:35:27 Unstable branch on crawl.develz.org updated to: 0.32-a0-1233-gb66479f5cc (34) 22:58:41 Windows builds of master branch on crawl.develz.org updated to: 0.32-a0-1234-gac1fde5492 23:17:05 <04dracoomega> I continue to be very much not a fan of Crawl's item handling backend 23:17:07 <04dracoomega> (Ranged shadow mimic ammo stuff is being a pain) 23:35:05 Unstable branch on cbro.berotato.org updated to: 0.32-a0-1234-gac1fde5492 (34) 23:55:36 Monster database of master branch on crawl.develz.org updated to: 0.32-a0-1234-gac1fde5492