00:01:02 the mysql container was still starting, you should see another attempt 5 secs later 00:01:44 you can check the container status with `docker-compose ps` 00:08:02 oh 00:08:29 no way to make docket wait/check for the service? 00:09:30 hrm, I keep getting those messages 00:10:18 docker-compose just says that loaddb.py is up 00:10:51 not able to connect to the web server either 00:11:22 I'm using test mode fwiw 00:12:56 hm. maybe the network isn't working properly 00:13:15 first thing I'd try to reset everything to good base state: docker-compose down -v. That will destroy the network 00:13:52 If it's still failing, maybe there is something wrong with docker network bridging on your machine? If you can't access the nginx server via localhost:8080 and it's up 00:14:22 in the docker-compose ps output, you should see port mappings like this for web & mysql: 00:14:24 dcss_tourney_mysql_1 docker-entrypoint.sh --def ... Up 0.0.0.0:3306->3306/tcp, 33060/tcp 00:14:26 dcss_tourney_web_1 nginx -g daemon off; Up 0.0.0.0:8080->80/tcp 00:15:47 ok, will try 00:21:16 alexjurkiewicz: ok I was just impation; I didn't realize it could take up to a couple of minutes for it to finish initializing 00:21:18 *impatient 00:21:44 it was easier to see what was happening when I added --force-recreate to the up command because that seemed to enable some verbose output 00:23:43 on an unrelated note 00:24:24 can't believe stuff like this is still in our user agents in 2020 Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/81.0.4044.138 Safari/537.36 00:25:10 it's chrome aka safari, based on apple web kit aka khtml which is like gecko 00:26:46 chrome is in the process of deleting user agent strings, so in 2021 it might finally be fixed 00:27:15 also wow couple of mins. I'm seeing about 15-20secs for first startup of mysql on my macOS laptop 00:27:29 (excluding image pull time, which was a few mins, if that's what you mean) 00:29:08 yeah it was a full 1-2 minutes from time of compose up excluding the container building part 00:35:00 Did Dr color get set during init/seed, like with other Race/Wr in 0.25? 00:44:18 no 00:48:06 advil was talking about doing that though 00:54:22 ebering: advil: FYI when I'm rewriting templates from old UI to new UI, I'm also moving all data loading from the mako file into render_page. IMO it's better for templates to be "dumb" both from a data separation PoV but also beause we can theoretically typecheck python code in .py easier than in .mako 01:24:17 I asked about it once in ##crawl and advil replied, but I wasn't sure if it was something that was considered a bug, or inconsistent somehow and if I should report/suggest it. I just figured it would be nice to play a seed and have the characters and maps as similar as possible for various reasons. 01:30:15 It seems a logical change to me 01:30:45 and I'm not sure what advil told you, but in Discord he suggested that he's not sure he likes the seeding of Ds either (even thought that's currently implemented) 01:31:32 it does mean that you can only have an experience as a given Ds/Dr with a specific seed and not a different experience 01:32:34 like if you quaff a mutation potion you'd not expect the seed to determine the results of the potion necessarilly 01:33:56 For weekly challenge, I'm happy with Demonspawn mutations being fixed per seed, and I would like Draconian to be fixed too 01:34:23 I have some reservations, but I think they boil down to "I feel like it lets people cheese demonspawn by finding seeds with optimal mutation combos". But that's their problem 01:36:23 well, seeded games are independent of any specific challenge 01:36:39 perhaps it's a thing that could be a command-line option (not enabled by default), for instance 01:36:59 it we decided to make ds mutations and dr colors independent by default 01:38:03 I can see the rationale for making them independent by default..almost seems like whether you want them independent depends on what you're doing with the seeded game 01:38:35 Unstable branch on crawl.develz.org updated to: 0.26-a0-8-gc4d20be10f (34) 01:38:44 I uh 01:38:55 didn't realize we'd changed trunk versions 01:39:38 those 0.25 branch tagged commits should have been an obvious sign 01:53:23 ebering: does it make sense to you to collapse all-players.mako and all-player-ranks.mako? 01:58:33 also suggestion for banner scoring: instead of 1,2,4 points, use 1,3,5 for greater spread 02:04:28 Ah, I was thinking the basic reason for seeds was to create reasonably similar games, and probably for debugging reasons. And does sound interesting to have a same-seeded game challenge. 02:05:50 it might be nice to use jinja2 instead of mako if possible, just for consistency with webtiles 02:06:35 not that jinja2 is especially great, but it'd be nice to standardise on something 02:08:50 yes, maybe for a future tournament 02:10:57 Windows builds of master branch on crawl.develz.org updated to: 0.26-a0-8-gc4d20be10f 02:10:57 oh, we use that for webtiles templates? 02:11:03 I vote we switch to PHP 02:11:09 react with an emoji if you disagree 02:11:25 👹 02:36:57 alexjurkiewicz: btw, from what I'm seeing, docker-compose up loaddb doesn't require docker-compose build loaddb beforehand, since the up action seems to imply a build 02:37:08 but correct me if I'm wrong there 02:37:39 it's a little confusing since my ubuntu docker-compose was too old so I had to update it to an install of the project's release 02:39:42 hrm, ok 02:41:25 alexjurkiewicz: so it seems that my previous problem with it never connecting was actually more related to the fact that mysql isn't properly starting if I specifically use 'up loaddb' 02:41:40 alexjurkiewicz: what I actually have to do is just use up with no arguments, that seems to properly start e.g. mysql and nginx 02:41:59 and startup was quite fast this time 02:42:35 but I'm not sure if this is due to differences between our versions of docker-compose or what; in either case I think that persistant mysql connection failure was because those other services simply weren't starting 02:43:15 ho boy, not sure if I love the new html style of this index page 02:44:59 ziggurnaut (L13 GrFi) Crash caused by signal #15: Terminated (Lair:6) 03:01:54 hrm, all my docker links have file:// urls 03:21:11 Unstable branch on crawl.beRotato.org updated to: 0.26-a0-8-gc4d20be (34) 03:37:37 fatjedi (L12 MiBe) Crash caused by signal #15: Terminated (Lair:2) 03:46:52 Fork (bcrawl) on crawl.kelbi.org updated to: 0.23-a0-3122-gcd7237df04 04:14:44 +gammafunk | but correct me if I'm wrong there <-- it won't rebuild if an image already exists 04:15:38 ah. Yes, the docker-compose file has no real sense of dependencies. You can add them, but I didn't bother. If you run `up loaddb`, it will only start that one container. You need to have `up mysql web` running in another terminal 04:16:09 the rules page is very rough. It's basically unstyled at this point 04:20:40 alexjurkiewicz: for some reason, the scripts seem to think that LOCAL_TEST is true despite that not being the case in the docker-compose.yml; I'm getting file:// links as if this variable is true 04:23:01 gammafunk: hm. One other uncommitted local change I have is `USE_TEST = True` 04:23:07 in test_data.py. Do you have that? 04:23:20 I set that to False 04:23:48 and did a rebuild 04:24:13 but I think LOCAL_TEST is independent of that 04:24:30 looks like the file:/// or not url logic doesn't care about it 04:24:47 the relevant pages still got generated, just had to manually browse to them 04:27:04 hm. I can't explain that. My first guess is the logic setting LOCAL_TEST is wrong 04:33:26 i extended it following the existing "misusing logic operators" probably incorrectly 04:36:57 oh, yeah 04:37:12 I saw that line of code and said "ug, I sort of remember this" 04:54:17 good type for an arg for an existing function: Optional[List[Union[Tuple[str,str],Tuple[str,str,bool]]]] 05:13:36 all HTML tables use new style now, so they will be ugly on old style pages. Check out `templates/exploration-order.mako` for an extremely simple example of new-style pages. Less typing! 05:37:58 does the code at https://github.com/crawl/crawl.git provide web play capacity? or just local builds like Linux, Windows, OSX? 05:40:44 Kireji: yes, it has the WebTiles server in the crawl-ref/source/webserver directory; see the README in that dir 05:41:08 thank you 05:41:58 looking at crawl/crawl/tree/master/crawl-ref/docs and don't see any docs like INSTALL, or where to start reading if you want to build 05:43:11 there's clearly a big team here already, don't mean to criticize 05:45:53 Kireji: https://github.com/crawl/crawl/blob/master/crawl-ref/INSTALL.md is install file 05:46:14 which we don't seem to link from the top level README 05:47:30 AHH, there it is 05:47:37 whew 06:10:28 Bananafish1988 (L14 VSEn) Crash caused by signal #6: Aborted (Orc:2) 07:01:25

Vehumet's headhunting guide

07:01:25

07:01:25 I: Kill any two uniques within two turns of each other. 07:01:25
07:01:25 II: Kill two medium or deep uniques within one turn of each other. 07:01:25
07:01:25 III: Kill two deep uniques on the same turn. 07:01:26 good old banner 07:04:03 -!- amalloy is now known as amalloy_ 07:09:41 ebering: to keep things consistent, if you could commit the values for test_data.py you are using locally that would be great. Then I can make sure to be testing the same dataset as you. (Or if you're using the current values in loaddb, I'll update test_data to match) 07:23:59 does the dev team collect/use stats or play history from online games to adjust game balance? 07:35:44 yes, with sequell 07:36:12 Kireji ^ 07:39:47 !lg 07:39:47 No games for Kireji. 07:42:12 :D 07:43:14 Kireji: that data is collected and we do look at it to see what players are doing, but to be clear we tend to avoid changing the game just because of these results 07:44:14 gammafunk: https://crawl.develz.org/info/index.php?q=foo know anything about why this is busted? 07:45:15 not sure actually, that was working a week or two ago 07:45:16 shalott's down too 07:45:28 https://loom.shalott.org/learndb.html is giving me a 404 07:45:30 perhaps Napkin change something that broke sqlite3 when he installed that instance of mysql for us 07:45:38 yes, that's Snark's server though 07:45:52 ah, doh, this is the shalott i've been hearing so much about 07:46:10 right, it's the server that runs sequell; not sure what's being changed there 07:46:22 he mentioned he's going to move sequell to a new server as he finds time 07:46:28 but I suspect he's not really finished that 07:46:47 it's possible that page being inaccessible is why the CDO page is now not working, not sure 07:50:44 i think the db is a better bet, just based on the error message 09:35:21 alexjurkiewicz: will do 12:05:39 re jinja2, tornado's built in (barely-documented) templating is just similar to jinja2, it's not actually using that 12:06:51 I *think* I basically prefer mako over jinja2 because it makes it easier to actually use python code 12:07:08 (but I've never really used jinja2 in a deep way tbh) 12:16:59 https://github.com/crawl/dcss_tv :eyes: 12:17:27 I guess this was on snark's account before or something? 12:17:38 i imported cause I went looking for footv before and couldn't find it 12:17:51 not actually planning to make any changes or anything 12:17:54 yep 12:18:48 alexjurkiewicz: for the banner order page the "banners earned" string comes out of the DB, since we have mysql 8 I could have the database coalesce that into a json object. Can python read a dict sanely from a json string? 12:19:18 that would allow a nicer display of "banners earned" 12:19:46 it can read a dict very sanely from a json string 12:20:28 sweet 12:20:35 although it's not really that hard to get a dict from any old query :D 12:20:50 well this query is aggregating things to compute a score 12:20:55 none of this fancy new 2000s technology 12:21:06 currently produces a string with GROUP_CONCAT 12:21:09 i'm no huge fan of jinja, i would just rather minimize the number of things i have to learn 12:21:53 jinja is different enough from mako that it would be some effort to do a conversion, you basically can't embed python in it 12:23:00 embedding python code into templates sounds like something that should be avoided if possible 12:23:14 ebering json.loads (though in the context of tornado there is tornado-specific calls for this) 12:23:34 but yeah, no need to convert anything 12:23:39 a very nice principled stand aidanh ;-) 12:24:16 gotta balance those principles with pragmatics at some point 12:24:47 preferably before crawlcode wrestles you and your utopian visions of consistency into the ground 12:24:47 to give you a sense for how it's used typically in our code, here's an example https://github.com/crawl/scoring/blob/master/templates/streaks.mako 12:24:57 does not seem particularly harmful to me 12:25:12 I agree that it could be bad 12:25:24 i dunno how you'd write a unit test for that 12:26:25 i guess extensive use of mock.patch.object 12:27:00 let me warn you that I personally have no plans to ever write unit tests for scoring 12:28:09 heh, me neither 12:28:39 partly because it's going to be a pain in the ass with the way these templates are written ;) 12:31:20 oh hm, it's not actually all that much code 12:31:43 yeah, the structure is usually, a single python block at the top to load some stuff in from the query module 12:31:56 I think if one converted this to jinja2 it would involve making that into a macro of some kind 12:32:22 mako supports passing a template a struct, so that block at the top can be moved to the calls to rendering 12:32:32 alexjurkiewicz: is doing some of that in the tourney repo 12:33:13 sounds plausible 12:33:49 because scoring is running on cao it isn't exactly a fresh version of mako, so I haven't tried to mess with the templates beyond the bare minimum 12:34:00 hard to tell what's supported 12:34:13 scoring is a batch job, right? 12:34:37 yes, it's basically a static site generator running in a loop if that's what you mean 12:34:45 same with tourney scripts 12:34:58 they originate from the same code at some point long ago 13:25:39 Unstable branch on crawl.akrasiac.org updated to: 0.26-a0-8-gc4d20be (34) 13:59:07 I wonder if we set `restart_after_save = true` would there be outcry? 13:59:20 it's much more "normal" behavior relative to modern games 13:59:35 s/set/defaulted 15:02:57 advil: iirc the original reason it stayed defaulted to false was just because it was kinda buggy when it was first implemented 15:03:50 i forget the details but there was some things that sometimes didn't reset properly maybe? but assuming that's no longer the case it'd probably make sense to change the default 15:04:02 there were* 15:04:07 ah quite possibly; restart_after_game has been true by default for a while so I think those issues have mostly been worked out 15:04:17 yeah, definitely there were many things that didn't reset properly 15:04:36 oh! sorry yeah i was thinking of restart_after_game actually, but maybe it's a similar case 15:05:55 yes, similar in terms of potential bugs for sure 15:06:13 when flipping the restart_after_game switch I do remember getting some communication from people who had a particular way of interacting with the crawl binary at the command line and didn't like the new 15:06:50 I think polymorph isn't working as a tmut miscast effect; even a short immobile polymorph can be very severe compared to the other miscast effects at slightly dangerous failure 15:07:10 what about removing the immobile forms? 15:07:41 from the miscast effect that is 15:10:39 03advil02 07* 0.26-a0-9-g750a858: Default `restart_after_save` to true 10(2 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/750a858e70e7 15:11:04 let's see if there's any protest 15:11:26 since the crawl community likes change so much 15:11:52 that's possible, but I also think that even at low power the form duration is too high 15:12:05 for a miscast 15:13:03 ah 15:13:15 is it possible to tweak it even lower? not sure if you've adjusted relative to poly power 15:13:29 I was thinking like ~1-2 turns on the low end when I made that suggestion 15:13:52 which iirc is not possible with a regular poly 15:14:44 yes, thats because all badforms use a base badform dur 15:14:47 of 150 auts 15:14:51 ah 15:15:22 and separating that out requires messing with player polymorph a lot more 15:15:39 in the original write-up I had tmut just dumping on extra contam 15:15:44 I think doing something cosmetic (or cosmetic plus a bit of damage) for the white color would be fine 15:15:46 ah or contam 15:17:09 Unstable branch on crawl.kelbi.org updated to: 0.26-a0-9-g750a858e70 (34) 15:17:49 maybe keep poly for worse ones and light contam for the second tier? (can never remember the terminology) 15:18:09 well they aren't tiered any more 15:18:34 this isn't for now but a new effect that randomly melded one slot with no upside might be nicely flavorful 15:19:04 makes your bones in the slot temporarily disappear or something 15:19:17 the "tiers" are just a quick status light for the relation between max miscast damage and current max hp 15:19:26 I see 15:21:22 well, contam would be a safe option for now in that it scales easily and involves well-tested code, maybe slightly boring 15:21:59 hm fr contam can cause poly at high levels 15:22:12 yeah thats a good one 15:22:20 for 0.26 15:27:54 03ebering02 07* 0.26-a0-10-g709782f: Revise transmutation miscasts 10(5 minutes ago, 2 files, 7+ 2-) 13https://github.com/crawl/crawl/commit/709782ff74ba 15:28:57 03advil02 {ebering} 07[stone_soup-0.25] * 0.25-b1-9-g4843cea: Default `restart_after_save` to true 10(21 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/4843cea77cc8 15:28:57 03ebering02 07[stone_soup-0.25] * 0.25-b1-10-gcef9adb: Revise transmutation miscasts 10(6 minutes ago, 2 files, 7+ 2-) 13https://github.com/crawl/crawl/commit/cef9adb76cc1 15:32:25 Unstable branch on crawl.kelbi.org updated to: 0.26-a0-10-g709782ff74 (34) 15:51:33 03ebering02 07* 0.26-a0-11-g9c5a184: Typo fix (Goratrix) 10(3 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/9c5a1843486b 15:52:10 03ebering02 07[stone_soup-0.25] * 0.25-b1-11-g213d97f: Typo fix (Goratrix) 10(3 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/213d97f8dacf 16:02:22 Unstable branch on crawl.kelbi.org updated to: 0.26-a0-11-g9c5a184348 (34) 16:25:59 03ebering02 07* 0.26-a0-12-g4aaa7df: More randart nouns (cebolla) 10(19 minutes ago, 1 file, 101+ 0-) 13https://github.com/crawl/crawl/commit/4aaa7df55cc7 16:25:59 03ebering02 07* 0.26-a0-13-gb168a4e: New randart apperance adjectives (cebolla) 10(7 minutes ago, 1 file, 14+ 0-) 13https://github.com/crawl/crawl/commit/b168a4ef397b 16:25:59 03RojjaCebolla02 {ebering} 07* 0.26-a0-14-g7b3c7a5: Add verbs to some randart names 10(8 days ago, 2 files, 93+ 4-) 13https://github.com/crawl/crawl/commit/7b3c7a5917ff 16:25:59 03RojjaCebolla02 {ebering} 07* 0.26-a0-15-g366e943: Add adjectives to the randart name generator 10(8 days ago, 2 files, 167+ 3-) 13https://github.com/crawl/crawl/commit/366e9434eded 16:26:55 03ebering02 07[stone_soup-0.25] * 0.25-b1-12-g2958ca1: More randart nouns (cebolla) 10(20 minutes ago, 1 file, 101+ 0-) 13https://github.com/crawl/crawl/commit/2958ca185404 16:26:55 03ebering02 07[stone_soup-0.25] * 0.25-b1-13-gc79aa09: New randart apperance adjectives (cebolla) 10(8 minutes ago, 1 file, 14+ 0-) 13https://github.com/crawl/crawl/commit/c79aa096b6e4 16:26:55 03RojjaCebolla02 {ebering} 07[stone_soup-0.25] * 0.25-b1-14-g05a5553: Add verbs to some randart names 10(8 days ago, 2 files, 93+ 4-) 13https://github.com/crawl/crawl/commit/05a55538a600 16:26:55 03RojjaCebolla02 {ebering} 07[stone_soup-0.25] * 0.25-b1-15-gc471598: Add adjectives to the randart name generator 10(8 days ago, 2 files, 167+ 3-) 13https://github.com/crawl/crawl/commit/c471598eb153 16:29:20 03Nikolai Lavsky02 {ebering} 07* 0.26-a0-16-g3fdfaa9: Display Zz and ? status icons for allies 10(8 days ago, 2 files, 5+ 2-) 13https://github.com/crawl/crawl/commit/3fdfaa9094b5 16:29:57 03Nikolai Lavsky02 {ebering} 07[stone_soup-0.25] * 0.25-b1-16-g62a4f28: Display Zz and ? status icons for allies 10(8 days ago, 2 files, 5+ 2-) 13https://github.com/crawl/crawl/commit/62a4f28727fb 16:30:16 -!- amalloy_ is now known as amalloy 16:32:21 Unstable branch on crawl.kelbi.org updated to: 0.26-a0-16-g3fdfaa9094 (34) 17:00:47 03ebering02 07* 0.26-a0-17-ge621825: Stimulate Xom on worship from a faded altar 10(6 minutes ago, 1 file, 3+ 0-) 13https://github.com/crawl/crawl/commit/e6218251f0ce 17:01:10 I'll let someone else make the call on whether to cherry-pick that 17:01:10 I think it's ok to cherry-pick but I can understand if there's an objection 17:06:52 seems fine to cherry pick 17:08:06 DocEon (L8 HOBe) Crash caused by signal #15: Terminated (D:6) 17:08:07 what does that actually do in practice? 17:08:13 !crashlog 17:09:12 http://crawl.berotato.org/crawl/morgue/DocEon/crash-DocEon-20200524-210805.txt 17:09:25 might just be someone disconnecting at the statup prompt 17:09:45 90s limit exceeded: killed !crashlog 17:17:36 Unstable branch on crawl.kelbi.org updated to: 0.26-a0-17-ge6218251f0 (34) 17:20:39 amuses xom 17:20:53 uh, twiddless the internal numbers of the random state 17:20:59 to be more favorable 17:21:01 a random amount 17:26:21 03ebering02 07[stone_soup-0.25] * 0.25-b1-17-ge1e08b9: Stimulate Xom on worship from a faded altar 10(31 minutes ago, 1 file, 3+ 0-) 13https://github.com/crawl/crawl/commit/e1e08b9fd0c5 17:26:51 03Byrel Mitchell02 {ebering} 07* 0.26-a0-18-g135ff06: Fixed terrain change duration not decrementing while off-floor 10(3 days ago, 1 file, 1+ 0-) 13https://github.com/crawl/crawl/commit/135ff066671a 17:32:12 Unstable branch on crawl.kelbi.org updated to: 0.26-a0-18-g135ff06667 (34) 17:35:59 big changes 17:36:49 fwiw, my preference for templates is to have no python code in them. I would rather run the python code in a python file and pass in primitive data types only 17:37:24 but yeah... that's not going to change before tourney. And I'm not even really fixing it as I go 17:41:06 following the compile instructions on Mac OSX, I finished "make -j4 TILES=y" and it seems to compile without error, but it's unclear where the binaries are. what am I missing? 17:42:00 Kireji: the binary would be in the source directory simply as "crawl" if you didn't specify the mac app target 17:42:20 there's a target to make an actual app, but there might be some invocations changes you need to build that 17:42:40 I'm not sure if advil has documented this for recent MacOs versions 17:43:04 how do I specify the mac app target? 17:43:37 probably belongs here https://github.com/crawl/crawl/blob/master/crawl-ref/INSTALL.md#macOS 17:46:21 reading Makefile 17:49:10 targets are mac-app-tiles and mac-app-console 17:49:43 it will write them into mac-app-zips 17:50:22 this reminds me. Are there any active devs who can notarise / sign the mac binary for 0.25.0? 17:53:27 ah, had to fix the python pyyaml thing, the previous build didn't go. redoing 18:00:53 \o/ 0.26-a0-16-g3fdfaa9094 working! 18:03:34 :) 18:03:39 The build passed. (stone_soup-0.25 - 62a4f28 #13470 : Nikolai Lavsky): https://travis-ci.org/crawl/crawl/builds/690713539 18:24:57 Unstable branch on underhound.eu updated to: 0.26-a0-18-g135ff06667 (34) 18:25:38 The build passed. (stone_soup-0.25 - e1e08b9 #13472 : Edgar A. Bering IV): https://travis-ci.org/crawl/crawl/builds/690728394 18:35:58 yeah, I'm not sure about those mac targets and whether you need special build setup for them to work 18:36:09 this is something advil has documented somewhere; maybe it's only relevant for release? 18:36:28 Do those github actions we're planning to use actually build the apps? 18:37:52 seems they do 18:38:29 so I'm wondering if those have been tested 18:38:29 maybe alexjurkiewicz has tested them since I believe he uses macos 18:39:01 alexjurkiewicz: to answer your earlier question about signing, geekosaur, a former channel regular, got this working for a couple of versions, but it was a nightmare for him and eventually broke and he wasn't able to get it working again 18:39:31 the signing has to be done as part of the build process, from what I understand, so not sure if it's even possible to integrate with those github actions, even assuming we could get it to work 18:53:01 The build passed. (master - 135ff06 #13473 : Byrel Mitchell): https://travis-ci.org/crawl/crawl/builds/690728485 18:59:56 i tested the github actions for release building in my own fork, seemed to work 19:00:26 but the real acid test is in the official rep 19:00:27 https://github.com/alexjurkiewicz/crawl-ref/releases/tag/0.50.13 19:00:30 alexjurkiewicz: as in the mac apps ran ok for you you? 19:00:49 I'm not talking about testing the system in general, just whether the macos app targets build from the github process work 19:01:11 if those default build args are fine, we probably should just add those to the INSTALL doc like Kireji suggests 19:01:25 and we should probably update the README to point to the install doc too 19:01:55 which is something I didn't realize we were missing; you'd only find the install doc after browsing into the unfortunate crawl-ref subdir we have to maintain 19:02:02 well, the macos binaries work, but only after an extremely scary warning from the OS which has only the option "move to bin" 19:02:19 oh, because they're unsigned? 19:02:32 I assume you get said scary warnings from the builds advil made on our download page as well? 19:02:36 right. You have to go into system preferences to unlock them 19:03:04 I think the install page will have to document the workaround, it's not intuitive unless you already know how 19:03:17 btw +1 to your install & readme suggestions. But I won't open a PR to do it :) 19:03:27 yeah, I can certainly make that change 19:03:47 my suspicion is that you have to do this step you're describing even with our previous official builds of the mac os apps (on the download page) 19:17:46 gammafunk: I just completed a build on 10.13.6 out of teh box from git, tiles works fine so far. the only thing I had to fix way pyyml 19:17:53 *the 19:18:03 *was 19:19:59 when I ran the executable from the crawl-ref/source dir with ./crawl, there was no warning 19:24:14 if you build it on your machine it runs fine 19:24:20 the issue is downloading precompiled binaries 20:11:46 Kireji: right, can you try to download the official macos binaries on http://crawl.develz.org/download.htm ? 20:11:49 like ebering says, you might get warnings about those 20:37:43 ebering: advil: I've got a working conversion of dcss_tourney to python3. How interested are we in using it for the tourney? 20:39:12 ooh, I was thinking of doing that myself 20:39:29 I don't think it would cause any problems on the CDO side of things; my concern was breaking anything people were working on 20:39:46 again, python 3.4 is what we have on CDO, so it can't be anything that requires a more recent version than that 20:40:05 I suppose we could ask napkin to install a more recent python3, but I'd like to avoid that if possible 20:40:48 my big concern is around silent bugs creeping in to the data. The conversion was pretty clean with 2to3, but I've had to rewrite a lot of sorting logic 20:41:16 possible; I guess we don't really have much of a testing infrastructure 20:41:42 we can run the scripts for a couple days with 0.25 (or 0.24) test data, but we'd have to spot any bugs ourselves 20:45:09 "a lot" = about four comparison functions 20:45:53 so actually I feel alright about it. The code is a little uglier because I've left in the dumb 2to3 changes mostly (eg lots of unnecessary `list(mydict.keys())` instead of `mydict.keys()` 20:47:59 alexjurkiewicz: I'm afeared of silent bugs 20:48:13 https://github.com/crawl/dcss_tourney/pull/7 20:48:27 also py3.4 v 3.7 issues 20:51:28 we can protect against that by testing with py3.4 locally 20:51:37 yeah, was going to ask if the docker could be updated 20:52:00 tbh the only change I'm really concerned about was rewriting a sort function in query.get_death_causes. Which is easy enough to test 20:52:43 I called out all code changes in the PR 20:54:51 Well I support a pre-tourney conversion, but will definitely defer to ebering, whose doing most of the scoring rework, as well as advil, who may be aware of technical issues I'm not 20:55:39 obviously if it doesn't get merged pre-tourney, would be good to do immediately after; we do have to test any conversion at some point 20:58:22 I'm not a python pro by any means, I'm trying to balance between using modern toolchain and not making more work than necessary 20:58:27 beware the feature creep! 20:58:48 but if alexjurkiewicz already has it running I'm happy to spot check it and merge it 21:01:07 i'll keep palying today and let you know later today if it's good 21:01:24 I have been wanting to gradually type a bunch of existing data to help refactoring 21:04:09 cebolla: how would you like to appear in CREDITS.txt 21:11:10 all these people I want to add to credits ducking and running 21:15:13 generate a pan lord name for each of them and use that 21:19:43 !commit Add Epicwang to CREDITS.txt 21:19:44 ebering: I don't know that we ducked and/or ran :-) 21:19:50 you did not 21:19:50 03ebering ⛐ 0.25-a0-2043-g881e100: Add Epicwang to CREDITS.txt 10(in the future, 28 files, 657+ 788-) 13http://s-z.org/neil/git/commit.png?p=crawl.git;a=commitdiff;h=881e100 21:23:44 does |amethyst's car crash emoji not render probably for everyone else? 21:23:50 I've notice it also does not render in discord 21:24:00 weird thing is it used to render properly 21:24:36 alexjurkiewicz: that patch is significantly less painful than I was expecting 21:24:53 ditto! 21:32:38 this may be another too invasive thing, but we really need to move the tourney config out of py files and into e.g. yaml config 21:32:50 maybe that's not too invasive and I can just be a good boy and go do that! 21:33:16 it's really a pain to test stuff when modifying files checked into the repo 21:34:52 03ebering02 07* 0.26-a0-19-g5d2c98b: Update CREDITS.txt 10(26 seconds ago, 1 file, 6+ 0-) 13https://github.com/crawl/crawl/commit/5d2c98bf94f4 21:35:28 03ebering02 07[stone_soup-0.25] * 0.25-b1-18-g5a45388: Update CREDITS.txt 10(62 seconds ago, 1 file, 6+ 0-) 13https://github.com/crawl/crawl/commit/5a45388c4366 21:36:31 that'd be nice yeah 21:36:46 if you could make USE_TEST read an env var it would basically be enough for me 21:45:23 heh. One issue with forcing a rebuild with docker-compose continually is that you'll eventually use up all your disk space with the rebuilt images 21:45:42 `docker image prune -f` will sort you out 21:47:15 alexjurkiewicz: random question for you, it seemed to me that the docker was putting the downloaded CAO milestones/logfiles somehow in my checkout's logfiles/milestones dirs when I'd thought that would all go in my system docker volume directories for that container 21:47:33 is the docker set up to duplicate those downloads locally, or did I just do something weird 21:47:54 specifically mean the files fetched by loaddb.py by wget 21:48:00 Unstable branch on crawl.kelbi.org updated to: 0.26-a0-19-g5d2c98bf94 (34) 21:48:15 these files in my checkout's subdirs are owned by root, so clearly made by the docker 21:50:03 the docker-compose.yml file mounts 'logfiles/' on the host to the loaddb container, and ditto for milestones 21:50:14 docker-compose will create empty dirs if the host mount target doesn't exist 21:50:38 are those so-called "bind" mounts then? 21:50:45 I couldn't see any defined in the docker-compose 21:50:57 I did this rather than storing them in the loaddb image so that you could update the code without needing to re-download the files. And I put the files on the host rather than a hidden docker volume because they'd be identical to when testing locally on the host 21:50:58 but good to know 21:51:03 yes, those are bind mounts 21:51:07 cool, thanks 21:51:18 see the services.loaddb.volumes section of docker-compose 21:51:23 and yeah that's probably a good call 21:52:14 same sort of logic for the webroot. Especially my work has involved manually editing html output and refreshing my local browser to prototype things 21:52:38 oh..didn't really those files were also bind mounted 21:52:47 now I wonder where the hell they are on my system 21:53:02 aha, just right there in the checkout 21:54:48 yup. The paths are relative to the docker-compose.yml location (eg if you are in templates, you can still run docker-compose commands and the paths will mount correctly) 21:56:59 thx 22:10:23 03gammafunk02 07* 0.26-a0-20-g5f92b90: Remove Lightli... 10(62 seconds ago, 1 file, 1+ 0-) 13https://github.com/crawl/crawl/commit/5f92b90fa0c7 22:11:05 03gammafunk02 07[stone_soup-0.25] * 0.25-b1-19-g693e237: Remove Lightli... 10(2 minutes ago, 1 file, 1+ 0-) 13https://github.com/crawl/crawl/commit/693e237038f4 22:14:32 ebering: I've pushed fixes for those sort order things, thanks for actually thinking about the ordering 8) 22:17:36 Unstable branch on crawl.kelbi.org updated to: 0.26-a0-20-g5f92b90fa0 (34) 22:23:14 this was all to be able to align numeric columns properly in tables 22:49:10 and now I can implement what I want without Py3, so let's backburner it 23:13:01 guess I don't need to worry about python3 yaml then, for now