01:46:10 <06mumra> Oh certainly and it only stuck out for me because I was looking at something similar messing around with a shop name generator which I need to finish at some point. do_mon_str_replacements also has a whole load of substitutions and none of them check for this either except for this one which happens to be the only one that is also one doing further getSpeakString lookups: // Replace with species specific insults. if 01:46:11 (msg.find("@species_insult_") != string::npos) { msg = replace_all(msg, "@species_insult_adj1@", _get_species_insult(foe_genus, "adj1")); msg = replace_all(msg, "@species_insult_adj2@", _get_species_insult(foe_genus, "adj2")); msg = replace_all(msg, "@species_insult_noun@", _get_species_insult(foe_genus, "noun")); } So 01:46:11 _get_species_insult also ends up calling getSpeakString (potentially twice) and certainly it seems jpeg thought this was expensive enough to justify checking the key first so she added this if -- but granted, this was 17 years ago, when some kind of performance/memory impact might have been rather more noticable 🙂 (The list of species insults is still quite a lot shorter than the colours list though) 01:46:12 <06mumra> Anyway it's rather easy to add some msg.find checks to avoid the lookup if needed, but I already got halfway through just moving the miscast messages to a txt as it's pretty straightforward and it makes that concern irrelevant either way (I'd be interested to see a real benchmark certainly, certainly in Javascript which is my usual language I know for a fact that doing repeated large string manipulations like this would be terrible, but 01:46:12 garbage collection is a whole different story compared to C++) In an ideal world everything would be moved to txt anyway so everything is translatable, but some things are quite a lot harder than others without maybe some small syntax additions to the txt format 03:17:03 How long does it generally take to compile a trunk on a typical server (without previous build files)? 03:17:10 I am using the update-trunk in the DGL script to build. Is this task inherently slow enough to take several minutes even on the latest CPUs (or is it possible that the script lacks options like -j to use parallel cores)? 03:23:11 Recommend using ccache. It takes a few minutes here. I recommend not building with -j, it sometimes smooshes up the tilesheets 03:32:28 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-5208-geafff8c3b6 03:37:44 <12ge0ff> Hmm, if I have ccache installed, would it make sense to build the game with -j, then do make clean and rebuild w/o -j? 03:38:36 <12ge0ff> (i had some graphics problems when building -j too) 03:38:43 Frankly I think that's just making work for yourself, it doesn't take _that_ long to build. cmou used to be a Raspberry Pi 3 and even there making it from clean was under the half hour 03:38:45 <12ge0ff> building *with 03:40:06 <12ge0ff> yeah, but with -j i get it compile in less than 10 min, compared to 33min in a single job 03:40:23 33 minutes! What on? 03:40:30 <12ge0ff> on a toaster 03:40:40 Yeah, but apparently a toastier toaster than a Pi 3 03:41:09 What issues in the build process could cause tiles to break when attempting parallel compilation? 03:41:47 *break = smooshes up 03:41:47 Makefile has bogus dependencies. But no-one's been able to find them 03:42:34 (partly because it doesn't seem worth looking very hard - and on an active server the last thing you want to do is tie up all the CPUs at once to build crawl...) 03:44:08 I am testing creating a container-based Crawl server, but the build time is terrible. It seems like it will take several hours to build all the game builds. I really need a way to reduce the initial build time... 03:44:58 If there are no issues other than tiles becoming smooth, I might try using the -j flag during testing. 03:45:02 You could maybe make (no -j) in rltiles then make -j in source 05:18:27 <06dolorous_84348> Good to know. And yes, it would be ideal to move everything to txt, but there are also so many code-specific bits. The Xom pseudo-miscasts that aren't in txt depend heavily on looking at the surrounding envirtonment, player inventory, etc. 05:30:40 Unstable branch on crawl.akrasiac.org updated to: 0.32-a0-1234-gac1fde5 (34) 06:07:38 @sentei Is there a reason why you're using GCC6 instead of the latest GCC during the build process? I thought compilers were backward compatible, but can't legacy forks be compiled with the latest GCC? 06:08:46 I am working on migrating the latest 6 commits of dgamelaunch-config to dgamelaunch-dcss-forks-server and updating the container version to 22.04. However, installing GCC-6 via apt is not working in the 22.04 container. Although there are many manual installation methods, they are not clean, and I don't like them. 06:09:46 Nonetheless, if there's no other way, I might have to consider manual installation... 06:10:28 Or sadly, I might have to keep using the 18.04 container. 06:11:30 <06mumra> Probably a lot of that kind of thing could be done with embedded Lua although it's not widely used yet to draw examples from Still there are tidbits in the miscast code like if (!mons_has_body(*target.as_monster())) msg = replace_all(msg, "'s body", ""); So even if someone translates the .txt they can't get around all the conjugations and pronouns and stuff like that that comes after. If we could pass in additional 06:11:30 context variables (so like passing in the monster so you can use it from lua) it would help but still be ugly (But also, I don't think it's realistic to imagine crawl ever being fully translated to another language, even if it was 100% supported ...) 06:16:03 About 10 to 15 years ago, the Korean Crawl community attempted a project to extract and translate all the strings from the code. However, they faced difficulties in understanding how each string should be translated. Although they successfully created fully translated builds for versions 10 and 12, this approach to translation was not sustainable 06:16:04 with new versions. 06:17:35 Therefore, I am considering translation method that focuses on the frontend output using regular expressions and parsing, rather than modifying the game core. With the recent advancements in LLMs, this task might become easier. 06:29:37 <09gammafunk> ASCIIPhilia: very strongly recommend just using the recent gcc or clang from ubuntu 22.04. There is no reason to use such old compiler versions like gcc6 for crawl, and doing so makes it harder to support recent c++ standards 06:31:25 tensugalgo (L17 DsAr) Crash caused by signal #6: Aborted (Shoals:1) 06:35:57 <06dolorous_84348> As long as there are decent examples of such embedded Lua; I don't know it offhand, but swiping bits of existing Lua and modifying them has worked for me in the past. 06:40:26 <06dolorous_84348> And any attempt to translate to a new language would require a lot of string overhauls; some have been done with the txt files. (Note that I haven't done any translations of anything, although I have worked with GNU gettext enough some years ago to know how its approach works; the overly conjugated stuff would have to be split up into separate strings, and that is... nontrivial is the formal way of putting it?) 06:57:11 <02sentei> ASCIIPhilia: no, compilers are not backward compatible for all programs, like the old versions of crawl which the forks are based on. I ran into the newer gcc throwing errors for a couple of the oldest forks, so I used the gcc6 for them specifically, other forks use gcc7. The forks would need some commits backported to be compilable on newer versions, which I didnt want to track down. I suggest removing those forks which can't be 06:57:11 compiled on the newer version, at least for the time being. 07:01:48 @gammafunk In the container setup supporting various forks created by sentei, we do install and use the latest GCC, but we also install GCC-6 separately for compiling some specific forks. I don't particularly want to use GCC-6 without a good reason, so I was curious if it was really necessary, which is why I asked him about it. 07:03:24 @sentei Editing the source to support the latest compiler sounds dreadful. I totally understand. Still, I want to support various forks for the sake of user experience. I’ll have to look for the cleanest way to do the manual installation. 07:04:11 <06mumra> Yeah with the way that a lot of strings are pieced together, and things like plurals, conjugations, pronouns are handled, it's pretty much impossible to translate everything in a way that makes sense even to a language with similar grammatical structure to English, let alone a completely different one! But it's a really interesting idea to just take this as a frontend problem and use LLMs, you might have something there... 07:06:47 It would be great if there were a flag in the latest GCC to compile according to the standards of legacy GCC versions. I don't know for sure, but I guess there's no such method, right? 07:08:05 <02sentei> Its possible I specifically may not have known the correct way to compile the old crawl code using a new compiler as I am not very experienced in C++ compilation. Perhaps someone more knowledgeable would be able to figure out the correct flags. 07:09:08 -std=whatever is supposed to do that 07:09:17 <09gammafunk> @ASCIIPhilia I see, sorry, I misunderstood 07:12:35 @mumra It doesn't seem to work perfectly now, but I previously translated the ability and spell windows into Korean using that method. I created a script that replaced strings received from the WebSocket and displayed them in Korean. As far as I remember, it worked almost perfectly from version 18 to around version 25. Even now, aside from a few 07:12:36 new skills and some UI string misalignments, it still seems to work. 07:12:36 https://gall.dcinside.com/board/view/?id=rlike&no=30364 07:12:37 https://gall.dcinside.com/board/view/?id=rlike&no=29282 07:12:37 https://gall.dcinside.com/board/view/?id=rlike&no=29133 07:12:38 https://gall.dcinside.com/board/view/?id=rlike&no=29705 07:13:23 It works quite well and seems to last a long time, but manually mapping the logic is extremely tedious. 07:16:20 @geekosaur Oh, I'll have to look into that. Thanks! 09:04:11 ASCIIPhilia: if you add bcrawl, gooncrawl, bcadrencrawl, or stoat soup, let me know and I'll add it to the scoring. Er, when I get home in a week or so and fix it. 09:07:18 -!- robin is now known as Guest3953 09:44:07 -!- robin is now known as Guest3562 09:48:42 -!- Guest3562 is now known as robin 09:53:22 03gammafunk02 07* 0.32-a0-1235-gd283c3be7c: feat: Add unarmed combat damage to clua 10(6 minutes ago, 1 file, 114+ 6-) 13https://github.com/crawl/crawl/commit/d283c3be7cf3 11:35:36 Pinkbeast I'll get in touch once the server setup is complete. Thanks! 12:15:46 <02Monkooky> next take on charlatan orb changes: boost skill levels by 1/5? of evocations skill level 12:16:02 @senteiI cloned your project (https://github.com/Rytisgit/dgamelaunch-dcss-forks-server) and then ran the following commands: 12:16:03 1. cd utils && ./build-testing-container.sh 12:16:03 2. cd testing-container && docker-compose up 12:16:04 However, I encountered the errors shown here: https://pastebin.com/rX2LiuUM 12:16:04 Am I missing something? 12:16:05 Note: I only modified the volume part of the docker-compose.yaml to create a new volume as shown here: https://pastebin.com/LMeAyk6c 12:23:46 <02sentei> Yeah, thats due to the wrong entrypoint, i created 3, 2 of which initialise the files like game binaries and logfiles, and the 3rd which just tries to start the server, so that everything doesnt get rebuilt 12:25:50 New branch created: pull/3825 (2 commits) 13https://github.com/crawl/crawl/pull/3825 12:25:51 03mumra02 07https://github.com/crawl/crawl/pull/3825 * 0.32-a0-1235-gdeb94c4331: Txt-ify miscast messages 10(42 minutes ago, 3 files, 456+ 264-) 13https://github.com/crawl/crawl/commit/deb94c433158 12:25:51 03mumra02 07https://github.com/crawl/crawl/pull/3825 * 0.32-a0-1236-ga760ce6c7a: Some tweaks and additions to miscast messages 10(42 minutes ago, 3 files, 101+ 65-) 13https://github.com/crawl/crawl/commit/a760ce6c7a50 12:27:43 <02sentei> ASCIIPhilia: I know it might not be right way to have done this, but I have been the only one using it so far. It should work if you replace the docker-entrypoint.sh with the docker-entrypoint-build-all.sh in the dockerfile 12:30:09 <02sentei> I think you might also be able to override it with a command in the docker compose as well 12:32:04 @sentei�docker-entrypoint-all.sh: Sets up everything. 12:32:05 docker-entrypoint-trunk.sh: Rebuilds only the trunk. 12:32:05 docker-entrypoint.sh: Runs the server using existing data without building anything. 12:32:06 Is this understanding correct? 12:32:06 Yes, it seems better to override it in the docker-compose file. 12:35:43 * docker-entrypoint-trunk.sh = docker-entrypoint-build-trunk.sh, -build- is missing 12:36:38 <02sentei> yes, the trunk build sets up all the directories and files for the server to use, so it's necesary the first time. Afterwards if you mount filesystem volumes, you don't need to rebuild on start since you would have the files saved in the corect state from the last time the server was running. Rebuild all also includes the install-crawl-versions.sh to rebuild all forks 13:01:03 @sentei Even after changing the entry point to docker-entrypoint-build-all, I encountered a similar error. So, I executed chmod -R 777 ./ at the root of the clone directory, rebuilt the Dockerfile, and ran the docker-compose command again. It seems that the build script started working properly after that. 13:01:05 I'm not sure if this is an issue with my environment or if there were files with incorrect permissions in the original git system. 13:06:25 <02sentei> likely incorrect permissions in git in that case, good that you got it working 13:08:23 This is not directly related to DCSS, but does anyone know of a Windows IRC client that allows for multiline messages and is not web-based? Web-based IRC clients disconnect every few hours, and while I have installed HexChat as an alternative, it is very inconvenient as it cannot handle multiline messages. 13:09:29 the IRC protocol doesn't really do multiline messages 13:10:54 it dates from the early 90s and was modeled on an IBM network protocol from the 60s/70s (BITNET Relay Chat) 13:11:05 I understand that, but on https://libera.chat, it seems to automatically split messages based on newlines. That's exactly the feature I need: the ability to automatically split and send multiline messages. 13:11:54 hm. hexchat can do that here but I'm on ubuntu. although I have to paste the message into it 13:12:00 but I may have something disabled 13:12:07 -!- robin__ is now known as robin 13:12:19 <09gammafunk> ASCIIPhilia: Do you use discord? 13:12:37 -!- robin is now known as Guest8555 13:13:14 @gammafunk Yes, I also use Discord. However, I heard that this channel is linked to a private Discord channel. 13:13:45 -!- Guest8555 is now known as robin 13:17:11 I heard that it's okay to ask questions in the DCSS channel of the roguelike server, but the conversations there seem too broad. This place seems better for focusing on the specific conversations I need. 13:18:53 @geekosaur As someone who used IRC in the early 2000s, I believe there is definitely a charm to IRC that modern chat protocols lack. It's inconvenient, but it has its own nostalgic appeal. 13:21:18 Anyway, it's still amazing to realize that IRC has been around since the 90s. 13:21:37 <02sentei> I looked changing the permissions in git, and it seems that git doesn't save them, so the correct way is to update the dockerfile to set them 13:22:22 <02sentei> I likely had also run the mass permission set up locally and forgot about it 13:24:30 @sentei Thank you for taking the time to check. For now, the goal is to ensure that it runs properly, but if necessary, I can modify it later to use appropriate permissions. 13:24:58 I think I started the build quite a while ago, but as expected, building all the forks is taking a lot of time. It's still in progress. 13:27:06 <02sentei> Yeah it's pretty slow, that's why I had the trunk only build to test locally. 13:29:14 oh, If I set docker-entrypoint-build-trunk.sh as the entrypoint and run it, and then set docker-entrypoint.sh as the entrypoint and run it, the server should run fine even though I won't be able to use other forks, right? 13:30:08 If that's correct, I'll do that when testing changes in the future. Good idea. 13:31:30 <02sentei> correct, but I think the way you edited the docker-compose file, it creates and destroyes the volumes automatically, so they wouldn't be saved. That's what the external parameter means on those, that you set them up beforehand so the data is saved 13:34:36 <02sentei> so you would need to run docker volume create versionsdb and 3 other volumes to have them saved in your filesystem 13:34:59 <02sentei> then it would work with the external tag 13:40:06 Isn't the volume deleted only when using the "docker-compose down -v" command with the explicit volume deletion flag? 13:40:12 Actually, I'm considering changing it to mount folders instead of using volumes, like this: 13:40:18 volumes: 13:40:18 - ./versionsdb/:/versionsdb/ 13:53:33 <02sentei> That should work as well, I don't remember the exact reason I set up the volumes this way. 13:54:58 <02sentei> Might have been that I was trying to look into ways of hosting the docker crawl server on a service which takes just a docker image and volume mounts, instead of a whole vm. 14:17:16 <09gammafunk> ASCIIPhilia: sent you a DM on discord 14:19:57 03mumra02 {dolorous} 07* 0.32-a0-1236-g9b32533982: Txt-ify miscast messages 10(3 hours ago, 3 files, 456+ 264-) 13https://github.com/crawl/crawl/commit/9b3253398286 14:19:57 03mumra02 {dolorous} 07* 0.32-a0-1237-g2ab4ef48e6: Some tweaks and additions to miscast messages 10(3 hours ago, 3 files, 101+ 65-) 13https://github.com/crawl/crawl/commit/2ab4ef48e6c9 14:19:57 03dolorous02 07* 0.32-a0-1238-g48138d7521: Fix spacing of a pattern name. 10(24 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/48138d75211a 14:19:57 03dolorous02 07* 0.32-a0-1239-g50be89a3ba: Adjust unseen alchemy miscast messages. 10(52 minutes ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/50be89a3ba63 14:19:57 03dolorous02 07* 0.32-a0-1240-gc7163d470e: Tweak wording of necromancy miscast message. 10(51 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/c7163d470e42 14:19:57 03dolorous02 07* 0.32-a0-1241-gc81a1b28c5: Add a few Xom messages referencing colour patterns. 10(2 hours ago, 1 file, 16+ 0-) 13https://github.com/crawl/crawl/commit/c81a1b28c588 14:24:37 <06mumra> nice 14:29:20 <06dolorous_84348> Good idea with the color patterns, by the way. 14:31:44 <06dolorous_84348> Also, testing the new Xom messages with crystal walls helped me find an unrelated bug involving redefined granite statues (and likely other things). 14:33:49 <06dolorous_84348> Which I just reported: Xom's pseudo-miscasts that mention terrain don't handle redefinitions, so you'll get messages about granite statues when only scintillating statues are in sight (since the latter are redefinitions of the former). 14:36:32 <04dracoomega> Xom sees things as they truly are. If the player is confused, well, that's just XOm 14:36:34 <04dracoomega> >.> 14:38:37 <06dolorous_84348> Indeed 🙂 14:50:51 03gammafunk02 07* 0.32-a0-1242-g08d867e968: feat: Unarmed attack ego description for clua 10(3 hours ago, 1 file, 24+ 0-) 13https://github.com/crawl/crawl/commit/08d867e968b9 15:14:52 <06mumra> Yes and I was thinking it might be appropriate for Xom as well, I was going to say so but clearly I didn't need to 🙂 15:15:40 <06dolorous_84348> I guess we both thought along similar lines 🙂 15:36:40 Unstable branch on underhound.eu updated to: 0.32-a0-1242-g08d867e968 (34) 15:45:50 <12aâ€‹sciiphilia> @sentei https://gist.github.com/refracta/d0ef293d18649d26cf7fd4a2e7d3fd6d#file-log-L8861 python [36mdlg-forks-server_1 |[0m File "/crawl-master/webserver/webtiles/server.py", line 821, in run [36mdlg-forks-server_1 |[0m asyncio.run(async_run_server(nonsecure_sockets, secure_sockets)) [36mdlg-forks-server_1 |[0m AttributeError: module 'asyncio' has no attribute 'run' Could this issue be related to the Python version? 15:45:50 I found that asyncio.run is available starting from Python 3.7. If it worked when you last worked on it about 3 months ago, the version should be correct. I'm not sure why this error is occurring. Has the Python version for the Crawl WebTiles server been updated recently? 15:49:27 03NormalPerson702 {GitHub} 07* 0.32-a0-1243-g30aeb729f7: Fix Hellfire Mortar power cap (#3820) (#3824) 10(30 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/30aeb729f792 15:52:05 <02sentei> if you updated the ubuntu version to 22, my first guess is that there now is some mismatch between the packages that are pulled in automatically, not very confident in that guess, but seems like it could be possible 15:53:02 <12asciiphilia> No, I'm using the unmodified version to verify if it works first. The migration work is being done separately. 15:53:38 <12asciiphilia> Therefore, it's version still 18.04. 15:54:35 <02sentei> that's weird indeed, maybe some package shifted the default version in these 3 months 15:55:33 <02sentei> I'm checking what I have on the current server that is running 15:58:41 <02sentei> looks like its python 3.6.9 on my server 16:02:52 <12asciiphilia> [docker-compose.yaml] entrypoint: ["/bin/bash", "-c", "python -V"] dlg-forks-server_1 | Python 3.6.9 It's same 16:03:42 I think recently (past few months) 3.6.x was desupported for Crawl? 16:04:13 07advil02 * 0.32-a0-751-g66263ccf9e: fix: update some CI python versions 10(9 weeks ago, 1 file, 11+ 9-) 13https://github.com/crawl/crawl/commit/66263ccf9e86 16:04:13 %git 66263cc 16:05:01 <12asciiphilia> 9 week ago XD 16:05:10 <02sentei> Yes seems like that would be it 16:06:25 <12asciiphilia> @sentei So, does the server you're currently running use a version of the Tornado backend code that predates that commit? 16:09:26 <12asciiphilia> Updating the Crawl core can be done without downtime, but updating the Tornado code will definitely require the server to stop. This seems inevitable, but automating this process could be a future task. (Or is there already a related script in the project?) 16:10:45 03dolorous02 07* 0.32-a0-1244-g16219f168b: Fix typo in alchemy miscast message. 10(80 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/16219f168b14 16:10:54 <02sentei> Yes that would be the case, I build the latest image in february. 16:24:57 <09gammafunk> We don't update python requirements frequently at all, because we have to coordinate any such update with multiple servers. We will definitely notify you if we're going to push a breaking change to webtiles where you have to update some library, so I wouldn't worry about e.g. updates to tornado versions 16:25:23 <09gammafunk> The fact that we have a bunch of official servers, some of which are running significantly older OSes than what you will be running will protect you 🙂 16:26:39 <09gammafunk> We will of course update python (and other dependencies) from time to time as things reach EOL, but we do this carefully and after coordinating with all server admins we can communicate with. So now that you're in this channel, we'll let you know about any future breaking change 16:30:03 <09gammafunk> And it is a very good idea to use our makefile webtiles core file update target to update webtiles when we push significant changes. Basically, your webtiles server will get core code updates automatically when we make changes to the webtiles codebase. This will never touch config files like config.py or yaml files that define your configuration. Some updates we make will require a WebTiles restart in order to get new functionality, 16:30:03 but those are not frequent. It's fine to only restart WebTiles when you see a commit changing something in WebTiles where you want that change now 16:30:26 <09gammafunk> I'm not sure how this interacts with the container, of course 16:30:44 <09gammafunk> But this is how it works for all the official servers using the normal install of dgamelaunch-config 16:33:00 <09gammafunk> (And just to be clear, all this webtiles updating happens as part of the trunk build update, thanks to the Makefile) 16:33:26 <02sentei> Yes, it works the same way in the container, when trunk build in run, the webtiles files get updated. What happened this time, is that I haven't restarted the webserver since february, so this passed me by, I was quite confused why the server.py has code in it that can't possibly run in the container, it's just that it hasn't needed to restart and was on the older version in memory 16:34:07 <09gammafunk> nice 16:34:22 <09gammafunk> sounds like ascii will be a-ok then 16:39:32 03DracoOmega02 07* 0.32-a0-1245-gce46c558ed: Don't cover stairs with dead hoarfrost cannons (particleface) 10(40 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/ce46c558ede8 16:46:30 <02sentei> @asciiphilia I updated the fork server config to point to my repo of crawl, where I reverted trunk to a bit before the change which broke this. I haven't tested, but it should work. I think this is the easiest way for you to test out before making bigger changes 17:33:41 <12asciiphilia> @gammafunk Thank you for the detailed explanation. @sentei https://github.com/Rytisgit/dgamelaunch-dcss-forks-server/commit/9acf9edeed88a8ba5dee318fa52b8be0bdbad9f6 I checked the commit. Thank you for your consideration. Upon closely inspecting the logs, I noticed that some builds encountered compilation errors. gnollcrawl [crawl-forks] https://gist.github.com/refracta/d0ef293d18649d26cf7fd4a2e7d3fd6d#file-log-L5522 gooncrawl 17:33:42 [upstream] https://gist.github.com/refracta/d0ef293d18649d26cf7fd4a2e7d3fd6d#file-log-L6638 xcrawl [crawl-forks] https://gist.github.com/refracta/d0ef293d18649d26cf7fd4a2e7d3fd6d#file-log-L7079 bcadrencrawl [crawl-forks] https://gist.github.com/refracta/d0ef293d18649d26cf7fd4a2e7d3fd6d#file-log-L8762 I understand that builds from upstream repositories might fail, but I am unsure why the crawl-forks builds failed. Do you have any idea what might be 17:33:42 causing these issues? 18:05:16 <02sentei> That is unexpected, best I can suggest for now is to clean out the data from the mounted volumes to see if anything changes, since that has helped me a few time before. Since there are still a few issues, I will try to set the docker up locally for myself tommorow, since I need to sleep now. 18:10:14 <12asciiphilia> Great. I'll continue working on this when I have time as well. Have a good night! 18:34:39 <05icoson> well, to be clear python does currently require a non eol version 18:34:51 <05icoson> which is py3.8+ 18:36:21 <05icoson> https://devguide.python.org/versions/ 18:38:28 <05icoson> %git 7c03013a13d 18:38:30 <04Cerebot> advil * 0.32-a0-23-g7c03013a13: feat: error when python is past eol (4 months ago, 1 file, 8+ 3-) https://github.com/crawl/crawl/commit/7c03013a13d3 18:39:59 <05icoson> historically we've supported ancient versions because of slow server updates, but this led to a pretty bad situation and I've tried to move us beyond that 21:45:02 03dolorous02 07* 0.32-a0-1246-g5069440637: Add another Xom message with colour patterns. 10(40 minutes ago, 1 file, 7+ 0-) 13https://github.com/crawl/crawl/commit/506944063726 21:48:33 04Build failed for 08master @ 50694406 06https://github.com/crawl/crawl/actions/runs/9168956343 21:49:19 03dolorous02 07* 0.32-a0-1247-g5f74a5512e: Add another pattern name. 10(2 minutes ago, 1 file, 2+ 0-) 13https://github.com/crawl/crawl/commit/5f74a5512eaf 22:35:17 Unstable branch on crawl.develz.org updated to: 0.32-a0-1247-g5f74a5512e (34) 22:58:05 Windows builds of master branch on crawl.develz.org updated to: 0.32-a0-1247-g5f74a5512e 23:13:53 Unstable branch on cbro.berotato.org updated to: 0.32-a0-1247-g5f74a5512e (34) 23:55:14 Monster database of master branch on crawl.develz.org updated to: 0.32-a0-1247-g5f74a5512e