00:03:29 03Aidan Holm02 07[randbook-test] * 0.25-a0-958-gde3eeb8: Extract function and add tests 10(12 hours ago, 4 files, 76+ 42-) 13https://github.com/crawl/crawl/commit/de3eeb8ea0d3 00:25:51 i'm happy to put my hand up for implementing pvs studio in github actions if you think it would be worth it 00:27:23 i was thinking about clang-tidy 00:28:40 although that's more like a style linter; i gather pvs is more vuln analysis 00:28:54 i have an old PR to enable clang-format for all crawl files 00:29:18 there was concern at the time about breaking git blame, maybe other concerns 00:31:02 would that work as a pre-commit hook? or as a set of auto-format rules in a file somewhere? 00:34:03 if it's all pre-commit, it won't break blame 00:35:06 well, you have to reformat all existing code to meet the style 00:35:41 clang-format can read a config file, so the idea is you can put the config file into the repo and editors will pick it up automatically 00:35:55 but the first time you edit every file it's gonna be big news 00:36:29 we can use the same approach we're doing for webtiles; incrementally enable style rules and files 00:38:46 btw is there a reason we still keep travis around? anything it covers that github actions don't? 00:39:24 kate-: I've tweaked the charms doc on the wiki; there's a tension between "don't proliferate wizardly items", "don't have a mix of wizardly and non wizardly items in a category" and "make charms available to all species" 00:39:46 the solution currently in the doc is to make a new item type: the charm, that goes in the amulet slot and is wizardly 00:39:54 which I will admit is a bit of a cop-out 01:03:43 https://www.youtube.com/watch?v=P05ONfLOqmY 01:04:35 aidanh: at the moment, gha doesn't announce to irc, I had to write a script to do that because nobody seems to have yet. I would do that if you're happy to delete travis.yml 01:06:18 sounds good to me as long as we're not losing any coverage 01:14:48 03Aidan Holm02 07* 0.25-a0-975-gcae8493: Extract function and add tests 10(13 hours ago, 4 files, 76+ 42-) 13https://github.com/crawl/crawl/commit/cae8493d645a 01:19:39 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-975-gcae8493d64 (34) 01:35:41 travis has a couple of build types that gha doesn't have, but they don't seem particularly important 01:38:41 Unstable branch on crawl.develz.org updated to: 0.25-a0-975-gcae8493d64 (34) 01:44:25 which ones? 01:48:40 BUILD_ALL=1, EXTERNAL_DEFINES=-UTOURNEY 01:48:44 those seem to be it now 01:50:31 i think we need at least one BUILD_ALL=1, not sure about the tourney one 01:52:19 \sure 02:04:30 New branch created: pull/1389 (2 commits) 13https://github.com/crawl/crawl/pull/1389 02:04:30 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1389 * 0.25-a0-976-g49bf755: Add a CI job for BUILD_ALL=1 10(7 minutes ago, 1 file, 5+ 0-) 13https://github.com/crawl/crawl/commit/49bf755111f7 02:04:30 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1389 * 0.25-a0-977-g7530b9b: Add build failure notification 10(6 minutes ago, 1 file, 19+ 0-) 13https://github.com/crawl/crawl/commit/7530b9b1b620 02:06:43 just a draft so far 02:07:49 actually, given I'll be pushing a lot of test crap i should probably leave it closed for now 02:09:57 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1389 * 0.25-a0-978-g2206e1f: fixup! Add a CI job for BUILD_ALL=1 10(3 minutes ago, 1 file, 11+ 1-) 13https://github.com/crawl/crawl/commit/2206e1f37bd4 02:09:57 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1389 * 0.25-a0-979-gd119213: fixup! Add build failure notification 10(3 minutes ago, 1 file, 1+ 0-) 13https://github.com/crawl/crawl/commit/d1192134ab4a 02:10:55 Windows builds of master branch on crawl.develz.org updated to: 0.25-a0-975-gcae8493d64 03:12:12 1) don't listen to vocal minority. 03:15:01 grunt_potential has tags that never execute 13https://crawl.develz.org/mantis/view.php?id=12252 by bcadren 03:33:32 how are messages like this 03:33:34 07:32:27 dcss-ci | Build failed for ci-notify-build-failure @ ae474d53 03:33:36 | https://github.com/alexjurkiewicz/crawl-ref/actions/runs/99808772 03:36:28 what, no pretty colours!? 03:37:14 looks fine to me; if it's a ci run for a pr, will the pr number / url be listed? 03:37:30 you won't get messages for PRs from this, only pushes to branches in the main repo 03:38:14 i opened a PR. There are some manual steps to merge it involving adding two secrets to the repo 03:40:37 that's a shame 03:40:59 added instructions to the PR OP 03:41:54 well, it'll work without that change. But every fork will send their build failures here 03:42:30 New branch created: pull/1390 (3 commits) 13https://github.com/crawl/crawl/pull/1390 03:42:30 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1390 * 0.25-a0-976-gb96886a: Add a CI job for BUILD_ALL=1 10(2 hours ago, 1 file, 16+ 1-) 13https://github.com/crawl/crawl/commit/b96886ac4421 03:42:30 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1390 * 0.25-a0-977-g1bdb6d5: Add build failure notification 10(2 hours ago, 1 file, 25+ 0-) 13https://github.com/crawl/crawl/commit/1bdb6d51d275 03:42:30 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1390 * 0.25-a0-978-g6d4ef98: Disable codecov data trimming 10(75 minutes ago, 1 file, 9+ 8-) 13https://github.com/crawl/crawl/commit/6d4ef98b5fe1 03:42:41 hm, ok, no thanks 03:42:41 heh 03:42:55 re codecov data trimming, could we just delete the code instead of commenting it out? 03:43:47 New branch created: ui-refactor (2 commits) 13https://github.com/crawl/crawl/tree/ui-refactor 03:43:47 03Aidan Holm02 07[ui-refactor] * 0.25-a0-976-gebd7756: Add key_recorder_raii helper class 10(57 minutes ago, 1 file, 15+ 0-) 13https://github.com/crawl/crawl/commit/ebd775692eb5 03:43:47 03Aidan Holm02 07[ui-refactor] * 0.25-a0-977-g252dd26: Decouple key input from targeting_behaviour 10(34 minutes ago, 3 files, 8+ 13-) 13https://github.com/crawl/crawl/commit/252dd269b3e8 03:44:12 i think there are faster ways to do the filtering than the commented out command. Even after spending a lot of time building coverage support, I still don't really understand how gcov/lcov ecosystem works. I figured it would be good to leave in as a prompt for someone smarter than me in future 03:44:50 reading that, I don't know what 'unimportant files' are 03:45:08 everything matching 'crawl-ref/source/catch2-tests/*' 03:45:29 oh so coverage for tests themselves? 03:45:47 yeah. I guess it's measuring how much dead code is in the unit test directory 03:45:57 seems like something we'd want to measure 03:46:10 It will contribute to the overall "coverage percentage" of the codebase now 03:46:23 hm, in that case I'll just delete the filtering entirely 03:47:09 anything to get our numbers up :) 03:47:34 no worries, i'll delete it myself when i merge this 03:47:46 since i have to merge it manually anyway to fiddle with the channel name 03:48:07 I've amended & pushed 03:48:13 had to rewrite the commit message anyway 03:48:44 ahh alright 03:49:18 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1390 * 0.25-a0-978-ga5352c5: Include coverage of catch2 files in coverage tracking 10(82 minutes ago, 1 file, 0+ 8-) 13https://github.com/crawl/crawl/commit/a5352c526ba5 03:50:36 so yeah, future improvements could be: a) more work to fully replace chei (you could build workflows to do so without too much work) b) support PR reports (status checks, new commits, open/close) in here 03:54:24 Fork (bcrawl) on crawl.kelbi.org updated to: 0.23-a0-3054-g1a545eca1e 03:56:26 chei can't quite be fully replaced, as we still need interactive commit lookup 04:09:09 true 04:12:17 hrm, directn.cc is impressively bad 04:14:55 yeah, directn.cc is a nightmare 04:15:12 my first crawlcode ever was working with that to make jump attack; I still have nightmares 04:16:45 i'm kind of considering just jamming a widget layer in the middle and fixing the bugs as they come in 04:16:45 this will all be in 0.26 trunk, of course 04:19:11 i think mouse click handling on console is definitely going to be broken for a bit, it's just too weird and disconnected from everything else to port cleanly 04:19:21 Fork (bcadrencrawl) on crawl.kelbi.org updated to: 0.22.1-2690-gd6882defc1 04:24:51 direction chooser weirdly contains various bits of description code that's needed for unrelated UI things, as I recall 04:25:15 but probably that's not the hard part of what you're planning to do 04:25:54 ah and that's not even in the class itself, just in that file 04:26:11 stuff like feature_description() is what I'm thinking of 04:26:40 should really be in that file, but that's an easy enough thing to change 04:26:56 s/should/shouldn't/ 04:27:38 that stuff is mostly fine, it's just clutter that obscures the 600-pound gorilla, i.e. direction_chooser 04:28:35 wow, TIL directn is short for direction, and not some weird "DirectN" thing related to ncurses 04:50:36 The build passed. (ui-refactor - 252dd26 #13246 : Aidan Holm): https://travis-ci.org/crawl/crawl/builds/684974984 04:55:07 03Aidan Holm02 07[ui-refactor] * 0.25-a0-978-g710f739: Remove switch fallthroughs 10(41 minutes ago, 1 file, 14+ 8-) 13https://github.com/crawl/crawl/commit/710f739ce224 04:55:07 03Aidan Holm02 07[ui-refactor] * 0.25-a0-979-ga611b47: Remove direction chooser show_items_once feature 10(10 minutes ago, 2 files, 1+ 7-) 13https://github.com/crawl/crawl/commit/a611b477ce53 04:59:24 03Aidan Holm02 07* 0.25-a0-976-gebd7756: Add key_recorder_raii helper class 10(2 hours ago, 1 file, 15+ 0-) 13https://github.com/crawl/crawl/commit/ebd775692eb5 04:59:24 03Alex Jurkiewicz02 {Aidan Holm} 07* 0.25-a0-977-gd920e8e: Add a CI job for BUILD_ALL=1 10(3 hours ago, 1 file, 16+ 1-) 13https://github.com/crawl/crawl/commit/d920e8e91e43 04:59:24 03Alex Jurkiewicz02 {Aidan Holm} 07* 0.25-a0-978-gbc0df25: Add build failure notification 10(3 hours ago, 1 file, 25+ 0-) 13https://github.com/crawl/crawl/commit/bc0df2583443 04:59:24 03Alex Jurkiewicz02 {Aidan Holm} 07* 0.25-a0-979-gd28c3ca: Include coverage of catch2 files in coverage tracking 10(3 hours ago, 1 file, 0+ 8-) 13https://github.com/crawl/crawl/commit/d28c3ca9de70 05:01:55 03Aidan Holm02 07[ui-refactor] * 0.25-a0-980-g8ea8630: Decouple key input from targeting_behaviour 10(2 hours ago, 3 files, 8+ 13-) 13https://github.com/crawl/crawl/commit/8ea86301a306 05:01:55 03Aidan Holm02 07[ui-refactor] * 0.25-a0-981-g07f64c5: Remove switch fallthroughs 10(48 minutes ago, 1 file, 14+ 8-) 13https://github.com/crawl/crawl/commit/07f64c5e9769 05:01:55 03Aidan Holm02 07[ui-refactor] * 0.25-a0-982-g49e7214: Remove direction chooser show_items_once feature 10(17 minutes ago, 2 files, 1+ 7-) 13https://github.com/crawl/crawl/commit/49e72149f428 05:02:47 one of my favorite classes in that mess is...I think it's called just "dist" 05:03:09 !source dist::isMe() 05:03:09 Can't find dist::isMe(). 05:03:19 and it's just a thing to pass arguments to direction_chooser 05:03:21 !source directn.cc:177 05:03:21 https://github.com/crawl/crawl/blob/master/crawl-ref/source/directn.cc#L177 05:03:37 why is it called dist?! and why does it exist 05:04:09 too long, needs urgent refactor renaming it to dst 05:04:43 renamed direction_chooser to dtncsr 05:07:38 it looks like it encapsulates output state for the direction chooser ui, like if a cell was chosen 05:07:59 which is definitely better than putting it all in direction_chooser, but still, the naming could be a lot better 05:08:47 right, the sort of confusing part is that it takes a dist as an argument in the constructor for direction_chooser 05:09:08 but maybe that's the best way 05:10:30 the best way is probably returning by value 05:10:44 i.e. output = direction_chooser(...).choose_direction() 05:15:21 The build has errored. (ui-refactor - a611b47 #13248 : Aidan Holm): https://travis-ci.org/crawl/crawl/builds/684986080 05:54:46 ebering, #1385 isn't just a flavour bug. There is such a thing like sack of spiders that can give you allied tarantellas at high Evo. People do use them to confuse oofs and pan/hell bosses. (Can't add note to the github issue now.) 06:07:35 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-979-gd28c3ca9de (34) 06:40:43 03Aidan Holm02 07* 0.25-a0-980-g8ea8630: Decouple key input from targeting_behaviour 10(4 hours ago, 3 files, 8+ 13-) 13https://github.com/crawl/crawl/commit/8ea86301a306 06:40:43 03Aidan Holm02 07* 0.25-a0-981-g07f64c5: Remove switch fallthroughs 10(2 hours ago, 1 file, 14+ 8-) 13https://github.com/crawl/crawl/commit/07f64c5e9769 06:40:43 03Aidan Holm02 07* 0.25-a0-982-g49e7214: Remove direction chooser show_items_once feature 10(2 hours ago, 2 files, 1+ 7-) 13https://github.com/crawl/crawl/commit/49e72149f428 06:41:19 03Aidan Holm02 07[ui-refactor] * 0.25-a0-983-gfd92ea8: Split directn command handling into separate method 10(12 minutes ago, 2 files, 31+ 21-) 13https://github.com/crawl/crawl/commit/fd92ea819522 06:41:19 03Aidan Holm02 07[ui-refactor] * 0.25-a0-984-gc76bb08: Consolidate directn shadow-step check 10(11 minutes ago, 2 files, 6+ 8-) 13https://github.com/crawl/crawl/commit/c76bb08d4f68 06:48:52 aidanh: thanks for merge. If someone could push a compile failure as a branch that would confirm the irc announcer is working 06:49:05 oh is it totally silent on success? 06:49:10 i hope so! 06:49:23 i'm not sure that's a good thing 06:49:36 building a 'notify on change' system would be quite difficult, if that's what you were thinking about 06:49:47 but if you wanted a message on every successful build, that would be possibl 06:50:13 the latter; something like travis's build passed notifications 06:50:47 usually i keep an eye on irc, and i don't really want to have to open up a web ui every so often just to see how my temp branches are doing 06:51:09 travis messages when the CI pass/fail state changes in a branch. It would be possible to do that with GHA 06:52:01 would need to add logic so that when a build completes, use the github api to go back through the branch history and find the most recent check status result. If it's different, send a message 06:53:39 I'll look into it, but might take a while. I haven't built exactly that before 06:53:51 i think sending a message for each build result is fine, no need to compare statuses 06:54:33 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-982-g49e72149f4 (34) 06:54:33 e.g. with this ui-refactor branch, i'd probably like to get periodic notifications that it's still passing 06:56:09 03Aidan Holm02 07[ui-refactor] * 0.25-a0-984-g2fbec62: Consolidate directn shadow-step check 10(26 minutes ago, 2 files, 6+ 8-) 13https://github.com/crawl/crawl/commit/2fbec62c1cdf 06:57:43 well whaddayaknow, the humble 'which direction?' prompt when you have multiple doors to open/close is actually directn itself! 07:02:14 alrighty, that's easy then! 07:04:15 thank you! 07:04:38 Build failed for ui-refactor @ c76bb08d https://github.com/crawl/crawl/actions/runs/99896037 07:06:36 awesome 07:13:49 gotta figure out why ccache isn't working, it's making the ci very slow 07:15:13 random question, do all jobs for the same project have a shared cmake pool? 07:20:52 is that to me about github actions? GHA has nothing to do with cmake 07:22:40 yes, and s/cmake/ccache/ 07:24:17 ah. It's configurable. At the moment, each job has its own cache. But there's a limit on total cache size which means I've set explicit max size for each job 07:36:20 ah. the reason CI is so slow is due to creating the mingw installer. It has to compile crawl twice 07:37:21 that does test the nsis install system, but nothing else special. I will remove it as the nsi file gets updated only biennially 07:42:25 03Aidan Holm02 07[ui-refactor] * 0.25-a0-985-g74c84bd: Split compass choosing out of directn.cc 10(8 minutes ago, 10 files, 151+ 110-) 13https://github.com/crawl/crawl/commit/74c84bdb6acc 07:44:08 03Aidan Holm02 07* 0.25-a0-983-gfd92ea8: Split directn command handling into separate method 10(74 minutes ago, 2 files, 31+ 21-) 13https://github.com/crawl/crawl/commit/fd92ea819522 07:44:08 03Aidan Holm02 07* 0.25-a0-984-g2fbec62: Consolidate directn shadow-step check 10(74 minutes ago, 2 files, 6+ 8-) 13https://github.com/crawl/crawl/commit/2fbec62c1cdf 07:54:09 aidanh: PR in to announce every build result. I'll be around tomorrow to troubleshoot if there are issues or further behaviour requests. PS the directn refactoring looks very cool based on my poor understanding of it from skimming the diff so far 07:54:28 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-984-g2fbec62c1c (34) 07:54:43 thanks and thanks! 07:54:47 i'll take a look at that now 07:55:10 New branch created: pull/1391 (2 commits) 13https://github.com/crawl/crawl/pull/1391 07:55:10 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1391 * 0.25-a0-983-g452b150: Notify IRC for every build result 10(31 minutes ago, 1 file, 46+ 8-) 13https://github.com/crawl/crawl/commit/452b150035e5 07:55:10 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1391 * 0.25-a0-984-gfff58cc: Don't build the Windows installer in CI 10(16 minutes ago, 1 file, 2+ 18-) 13https://github.com/crawl/crawl/commit/fff58cc5ea1d 07:55:39 well, i'm mostly just trying to disentangle the behaviour and logic from the ui control flow 07:55:55 to do that, I need to whittle down the amount of cruft in there 08:00:21 kennysheep_nondecorative_pillar has a depth error 13https://crawl.develz.org/mantis/view.php?id=12253 by bcadren 08:02:08 03Alex Jurkiewicz02 {Aidan Holm} 07* 0.25-a0-985-gf7bcf91: Notify IRC for every build result 10(38 minutes ago, 1 file, 46+ 8-) 13https://github.com/crawl/crawl/commit/f7bcf9115c59 08:02:36 alexjurkiewicz: i've just cherry-picked the notification change for now; i'm curious if there's some other way to improve the nsis test 08:03:22 alexjurkiewicz: e.g. perhaps if all jobs can share ccache, it won't matter if it's built twice? 08:03:33 notify on every build result seems maybe a bit much? 08:04:27 btw travis is already doing the secret thing and we *definitely* want that 08:05:01 it might be, i guess; not sure how noisy it'll be in practice 08:05:04 the secret thing? 08:05:18 without it practically everyone who forks the branch ends up spamming this channel at least once, and sometimes they're hard to contact 08:06:06 oh, right; that's already set up for gha irc notifications 08:17:30 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-985-gf7bcf9115c (34) 08:18:14 03Aidan Holm02 07* 0.25-a0-986-g94ad2f9: Split compass choosing out of directn.cc 10(44 minutes ago, 10 files, 151+ 110-) 13https://github.com/crawl/crawl/commit/94ad2f9d5aee 08:34:59 Build passed for master @ f7bcf911 https://github.com/crawl/crawl/actions/runs/99931435 08:38:00 Build passed for master @ 94ad2f9d https://github.com/crawl/crawl/actions/runs/99944532 08:38:04 haha 08:38:05 so 08:38:07 maybe not 08:40:33 03advil02 07* 0.25-a0-987-gc6157ca: I'm gonna veto this one, sorry 10(38 seconds ago, 1 file, 0+ 2-) 13https://github.com/crawl/crawl/commit/c6157ca55217 08:41:11 :( 08:41:25 does your irc client not filter joins/parts? 08:41:36 well, I could set it to do so 08:41:55 <|amethyst> I don't suppose there's a way to track changes of state there? 08:42:02 yeah, change would be fine 08:42:09 or once a day or something 08:42:18 or in a different channel 08:42:51 I wonder if my commit actually succeeded in vetoing that 08:43:55 <|amethyst> advil: probably not 08:44:00 I think even with joins/parts filtered that would be much too spammy for my tastes 08:44:27 <|amethyst> sounds like it would just print "for master @ deadbeef https://..." now 08:44:47 <|amethyst> maybe move the echo into the if? 08:46:21 <|amethyst> oh, no, I guess that wouldn't do it 08:46:32 03advil02 07* 0.25-a0-988-g4617b10: Probably actually don't message on every push 10(2 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/4617b1041bba 08:46:52 I changed the action condition back to what it was before 08:47:28 aidanh also I'm guessing you have email notifications off? 08:47:32 i do 08:47:37 because I'm already getting emails for all of this too :-) 08:49:09 https://github.community/t5/GitHub-Actions/Send-success-notification-after-failure/td-p/32404 08:49:19 hopefully at some point they will just add e.g. a "recovered" check 08:51:32 github does seem to be making many improvements these days, so it seems reasonably likely 08:53:22 it seems hard but maybe not impossible to implement right now, I think in principle you can use their workflow api to query the status of arbitrary workflows 08:53:52 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-986-g94ad2f9d5a (34) 08:54:23 i think i can cope with no recovery notifications if it means we can ditch travis 08:57:46 maybe something would be possible with https://developer.github.com/v3/repos/statuses/ 08:58:28 esp. if github actions is already setting statuses? 09:02:39 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-988-g4617b1041b (34) 09:03:14 haha I wonder if you could get the bot to PM the commit author 09:04:29 you'd need an email -> nick map 09:04:42 that we have 09:04:55 what about just doing these in another channel? 09:05:15 that's probably easiest 09:05:47 that's how it would be set up anyways if we were using slack integrations or w/e like good technocapitalists 09:05:59 man, how does this keymap context stuff even work at all 09:06:09 barely 09:07:23 like, what is it even supposed to do? and the whole 'keys are positive ints, commands are negative ints' is insane 09:08:40 I think it's basically to allow macros to be specific to certain contexts 09:09:56 where part of that is really *preventing* macros from working outside of the context they're defined in 09:10:54 well, I guess macros and bindings in general 09:11:56 wonder what happens if you switch contexts halfway through a macro... 09:12:12 for master @ c6157ca5 https://github.com/crawl/crawl/actions/runs/99952171 09:12:27 heh, mystery statuses; much better 09:12:32 haha 09:14:08 I think travis also (inconsistently) does some smarter agglomeration of messaging, it won't always send messages/emails when there are still queued jobs (though I'm not sure I really understand how it decides) 09:14:11 !source fire_target_behaviour::get_command 09:14:11 1/1. https://github.com/crawl/crawl/blob/master/crawl-ref/source/throw.cc#L242 09:14:55 so iiuc, that hack could be avoided by adding CMD_TARGET_* commands that are mapped to the same key? 09:16:14 yeah 09:16:15 %git 991379fdf84cc864 09:16:16 07wheals02 * 0.22-a0-124-g991379f: Make i, (, and ) in the targetting UI obey keymapping. 10(2 years, 3 months ago, 1 file, 19+ 9-) 13https://github.com/crawl/crawl/commit/991379fdf84c 09:16:27 that commit message gives an explanation of sorts for why that wasn't done 09:17:11 that's what my oblique comment about "if commands are really supposed to work in multiple contexts" is about 09:18:36 ah, right 09:19:08 i think i'd rather just copy the bindings over when loading keymaps 09:19:45 also, of course a function called get_command is actually *executing* the command and then returning 'no command' 09:19:47 totally demented 09:22:22 <|amethyst> aidanh: you are now immortalized in twitter 09:22:26 <|amethyst> <+aidanh> also, of course a function called get_command is actually *executing* the command and 09:22:29 <|amethyst> doh 09:22:32 <|amethyst> https://twitter.com/crawlcode/status/1259111312921067526 09:24:07 %git 0eb7e87096f0750 09:24:07 07greensnark02 * 0.3-a0-652-g0eb7e87: [1746014] Added an implicit quiver when firing missiles: - ^P/^N change the selected missile, but only within a fire_order slot. i.e., if you're wielding a launcher and have "launcher" first in your fire order, ^N and ^P will cycle only through suitable launcher ammo. - If you use ^P or ^N to change the default missiles offered, that choice gets saved as your quiver (unless you cancel targeting). - Fire always offers the quivered item first, if appropriate (so if you have darts quivered and try to shoot from a bow, the darts will not be offered). Breaks save compatibility (quiver is saved). 10(13 years ago, 8 files, 319+ 103-) 13https://github.com/crawl/crawl/commit/0eb7e87096f0 09:24:21 apparently the original design! 09:24:35 sry for the ping 09:26:01 that looks rather clean actually 09:27:29 yeah it has rather morphed over time 09:28:40 # and resides in des/builder/shops.des. 09:29:59 That comment at the top of bazaar.des doesn't appear to be true anymore. At least I can't find any functions related to bazaars outside of bazaar.des and none of the functions at the top of bazaar.des relate to the odds of the portals spawning. Or am I missing something? (I literally grepped the word bazaar so it's entirely possible I'm missing 09:29:59 something) 09:31:17 |amethyst: :) 09:31:31 <|amethyst> Bcadren: there's something there, see serial_shops 09:31:46 <|amethyst> Bcadren: search for bzr_entry 09:33:02 |amethyst thanks; it's literally in the "place a generic shop" vault then; seems like an odd place for it; but...thanks. haha. 09:33:58 <|amethyst> Bcadren: the "place (potentially) several shops" vault, but yeah 09:34:59 Oh I get it; it's a vault that only places subvaults; that's all it does. I'm tired. >_< hah. 09:35:15 03Aidan Holm02 07[ui-refactor] * 0.25-a0-987-gbd4ef80: Remove duplication 10(67 minutes ago, 1 file, 3+ 4-) 13https://github.com/crawl/crawl/commit/bd4ef80ceaf0 09:35:15 03Aidan Holm02 07[ui-refactor] * 0.25-a0-988-g27d5232: Fix ugly code 10(26 minutes ago, 1 file, 3+ 3-) 13https://github.com/crawl/crawl/commit/27d523286862 09:38:33 I've been reading vaults for the last ...idk 10 hours or so...I really should sleep. 09:38:46 <|amethyst> and seek therapy 09:42:41 aidanh: since you're working on targeter ui stuff, I dunno if you read all the way to the bottom of the UI improvements but a desired feature is to allow ^x from the targeting prompt, have it display only in-range applicable targets, and have selecting move the targeter to there. this comes from our screenreader players 09:48:12 that seems doable, but it'll have to wait for all the refactoring work 09:48:46 of course 09:49:44 I'm not sure whaat your refactoring notes are and if your plans will make that easier or harder or could do either dependinng on decisions you make 09:49:59 this was a request that you pick the "easier to do that" option :) 09:52:38 ...therapy? O_O; 09:55:04 Build passed for ui-refactor @ 27d52328 https://github.com/crawl/crawl/actions/runs/99977753 09:55:20 <|amethyst> Bcadren: extended exposure to crawl vault syntax has been clinically shown to lead to deep-seated rage and self-loathing 09:58:43 Something I made awhile back specifically as a tweak for gauntlets keeps being useful in various vaults. If y'all would like the idea for mainline. Runed_Clear_Stone_Wall ...player sees through it, monsters don't...pretty useful for when you want to play with player knowing the risk before opening a door, etc. (I also made runed clear doors work 09:58:43 the same)...like I said the original use for it was gauntlets so the player could view all the areas ahead of time without sacrificing the possibility of trying stealth. 09:58:43 The build passed. (master - 4617b10 #13260 : advil): https://travis-ci.org/crawl/crawl/builds/685026480 09:59:52 <|amethyst> huh, neat. I'm a little surprised asymmetric LOS didn't uncover an avalanche of assumptions in the code 10:01:38 |amethyst thing is; asymmetric LOS already exists... Ashenzari's Scrying. :) 10:01:43 <|amethyst> true 10:02:17 <|amethyst> also, I think I'm a jerk, because my second thoughy was "you should totally have the inverse of that too" 10:02:47 <|amethyst> s/thoughy/thought/ 10:03:12 monster you can't see smiting you from behind a grate? *shrug* IDK it could be interesting. 10:03:54 Also IIRC; monsters that have digging have their AI set to treat rock walls as clear; which is another case of asymmetric, but in the monsters favor. 10:05:49 <|amethyst> I think that only affects movement, not sight, at least in vanilla 10:08:35 yea...it's been months since I've been in that section of the code. The runed clear walls are 3 months old now; and me looking at how digging monsters move would have been when I restored 0.14 and earlier behavior of letting insubstantials like wraiths go through walls, etc. 10:09:56 <|amethyst> huh, actually, looking into that I did notice something funny 10:10:10 should be pretty much the same difficulty, since it'll probably just be adding a case in the direction_chooser command switch statement 10:10:37 <|amethyst> monsters with M_BURROWS or the digging spell will beeline straight towards you if you're their target (they just "know" your location without having to see you, but only when plotting a move) 10:10:58 <|amethyst> *but* that doesn't apply to monsters with wands of digging 10:10:58 i just really want to get this thing merged and done with before I totally forget how it all works :) 10:11:56 that sounds right. 10:15:49 it's basically a remnant of rock worms and aforementioned insubstantial behavior; it makes sense for things that move through rock to path through it accurately; a goblin with a wand of digging on the other hand lmfao. 10:16:29 <|amethyst> but DTEM and Jorgrun? 10:16:55 *shrugs* also Cacos have dig. 10:17:20 <|amethyst> Yeah, Cacodemons and Mnoleg I could kind of see because it's a "magical ability" 10:17:31 <|amethyst> but for DTEM and Jorgrun it's a wizardly spell 10:18:30 make Jorgrun worship Ash. hah. 10:21:29 the alternative would be treat the player as invisible and attempt to target dig with some fuzz? 10:22:48 <|amethyst> I'd be inclined to just treat goblins with /dig the same way as spellcasters 10:27:47 -!- Amnesiac_ is now known as Amnesiac 10:27:52 movement target as if it was trying to move closer to an invisible target if it picks to move to a rock wall; fuzz the vector of dig a bit? would the end result be that practically different? 10:45:35 <|amethyst> Bcadren: no fuzzing, just a preternatural ability to beeline, same as other diggers 10:50:10 So no change (yea I agree it's not worth the trouble; but I was just spitballing what a more 'flavourful' would be like since "it sees through walls only when planning a move" is indeed a bit odd). 10:50:59 <|amethyst> no change to casters, but right now wand users don't do that. mostly I'd want to get rid of the inconsistency, and also simplify the code a bit 10:53:09 ahh. Ok. wands of digging are long gone in my fork; moved it back to being a spell, but level 6 instead of 4 so it would take more investment to use it to make killholes in heavy armour. 11:02:44 aidanh: cool 11:18:39 Ok all: I've tweaked the tournament scoring proposal based on player input and various other feedback collected 11:19:06 https://github.com/crawl/crawl/wiki/New-Tournament-Scoring 11:19:06 One thing that's cosmetic that isn't in there 11:19:39 is to instead of using harmonic mean, use 10,000 / HM for display because watching your 8th decimal place go down 2 to pull ahead is not as exciting as earning MEGA POINTSS 11:24:35 I would like to start implementing next week so I want to make sure there's broad buy-in 11:37:16 this looks good to me 11:37:36 I wonder if there are more non-winner things that could be added? 11:37:59 a suggestion from reddit was for more non-winner small arity categories 11:38:13 the "Winning" category is modeled on it 11:39:38 LGTM, my only concern is how intuitive the ranking will be for players to interpret 11:39:49 where it = the idea of small n-arity categories 11:39:50 that looks like a good category 11:40:06 I did get some feedback about this that it didn't look very different for people who are aiming for 0-1 wins, I'm not sure that's entirely true but just in the raw category list it may look that way (and that winning category wasn't there at the time I think) 11:40:54 yeah, the winning category is new, and aimed at giving those 0-2 win players a dramatic rating bump 11:40:54 maybe it would be worth sorting the relevant categories together 11:40:54 when they get a win 11:40:54 yeah 11:41:33 I think that is a very nice idea 11:45:14 would you apply softmaxes to any of these? 11:45:38 ranking does away with some of the need to that but it could still mitigate some grinding on the high end 11:45:46 *need for that 11:49:08 Pious seems redundant with the piety category? 11:54:01 Yeah, some of the banners are redundant, that's a bit on purpose 12:07:13 ok, I updated it to group relevant categories 12:38:49 https://www.youtube.com/watch?v=oz5JQ_vJ6Dk 12:41:51 this is amazing 13:07:02 Odd display issue on CAO right now with the quiver showing empty. https://imgur.com/ivtyDet.png 13:07:26 Not sure if it's a display issue really, that's just what it looks like right now. 13:07:39 Unable to show quiver stones in display. 13https://crawl.develz.org/mantis/view.php?id=12254 by Euph0ria 13:24:49 wow, that kept getting better and better 13:25:18 I thought nothing would top the kraken until I saw the bouncing big kobold heads 13:31:37 that bug is probably due to one of my recent directn commits 13:32:19 this doesn't look too urgent, but feel free to revert if you want 13:34:49 yeah the kobold heads were good too 13:35:19 the kraken is my fav though 14:14:43 03advil02 07[stone_soup-0.24] * 0.24.1-9-ga107495: Standardize conditions for including unistd.h 10(25 hours ago, 14 files, 23+ 12-) 13https://github.com/crawl/crawl/commit/a107495faca6 14:14:43 03advil02 07[stone_soup-0.24] * 0.24.1-10-g7182e98: Try to fix cross-compiling 10(23 hours ago, 12 files, 12+ 12-) 13https://github.com/crawl/crawl/commit/7182e9863bd8 14:14:43 03advil02 07[stone_soup-0.24] * 0.24.1-11-gbbbb219: missed one 10(22 hours ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/bbbb2193e69f 14:21:30 Euph0ria: I haven't yet replicated yet replicated that quivering issue, either locally or on cao, can you point me to the relevant morgue? 14:28:45 also I *think* the report is about 0.24 so you're probably off the hook aidanh 14:32:32 ah got it, it's a dup of https://crawl.develz.org/mantis/view.php?id=12117 14:32:46 longstanding webtiles-specific issue if the ammo is on a 15:04:16 03ebering02 07* 0.25-a0-989-gc08aec6: Unstack tremorstones and make them indestructible 10(27 minutes ago, 5 files, 10+ 11-) 13https://github.com/crawl/crawl/commit/c08aec66ebdc 15:47:08 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-989-gc08aec66eb (34) 16:15:02 With Ru's sacrifice of courage, "_You now exude an aura of power that intimidates your foes." 13https://crawl.develz.org/mantis/view.php?id=12255 by RoGGa 16:39:17 03advil02 07* 0.25-a0-990-g0adbbaa: Fully sync player on webtiles load (12117) 10(10 minutes ago, 2 files, 12+ 0-) 13https://github.com/crawl/crawl/commit/0adbbaad2a99 16:48:33 advil: did you happen to see my messages to you the other day about changing scroll of aquirement behaviour to make it cancelable? 16:49:12 oh 16:49:25 wondering specifically if our 'reserving' the unrand until the player actually make their acquirement selection would affect seed stability 16:49:40 s/the unrand/any unrand generated in acquirement choices/ 16:50:03 with respect to seeding, I don't it would be any different 16:50:31 does acquirement actually generate / keep around the item? 16:51:04 if not there would be a tricky case where it provides an unrand, and the player triggers a level change that generates the unrand, and then they come back to the acquirement list 16:51:05 it currently generates the item_def but ensures that the unrand seen state is unseen 16:51:24 right, I was thinking the scroll would actually reserve the unrand until the player made their selection 16:51:30 -!- amalloy_ is now known as amalloy 16:51:50 yes, that would be fine for seeding if you can get it to work 16:51:52 but tbf that might require a little refactoring of how unrand seen states are tracked, would need to look into that 16:51:56 yeah 16:51:57 cool, thanks 16:52:22 it doesn't really need to have been seen but it needs to get added to the generated list 16:52:51 ok, maybe those are seperate lists then; either way I'll take a look at it 16:53:13 it's either that or keep adding every possible crawl info screen as a hotkey to the acquirement screen, since players keep requesting those 16:54:03 ah I see 16:54:10 what about just ?# 16:54:13 that covers everything 16:54:24 not stash searching! 16:54:28 haha ok 17:00:44 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-990-g0adbbaad2a (34) 17:07:32 I'm staring at 12255 and not seeing the bug 17:07:41 is it just that the flavor of the msg combination is weird? 17:10:41 I wondered about that one, IDGI 17:34:54 <|amethyst> yeah, I think it's supposed to be the flavour 17:35:35 <|amethyst> which isn't a great argument, because it's entirely possible to give off an aura of intimidation despite being a coward 17:35:38 <|amethyst> just ask my students 17:48:31 BUILD_ALL job on GHA isn't actually building all. I'll look into it 17:51:31 -!- aidanh_ is now known as aidanh 18:20:15 -!- Raichvent_ is now known as Raichvent 18:23:14 |amethyst: speaking as a student, that's what we always suspected of professors, but we were of course too cowardly to say it 18:23:37 <|amethyst> don't worry, we're intimidated by you too 18:23:52 Unstable branch on underhound.eu updated to: 0.25-a0-990-g0adbbaad2a (34) 18:29:36 |amethyst: hey neil, long time no see! 18:29:43 <|amethyst> bh: howdy 18:29:53 <|amethyst> classes are over for the semester 18:30:06 <|amethyst> https://i.redd.it/tuu6xfgknsx41.jpg 18:43:48 if they refactor beam.cc surely that's worth credit 18:47:33 bolt beem; has got to be some of my favorite crawlcode 18:48:20 <|amethyst> has that been posted to the twitten? 18:48:42 <|amethyst> s/ten/ter/ 18:49:12 dunno. setup_targetting_beam ... bolt beem; it's beautiful 18:49:15 who runs the twitter account? 18:49:31 <|amethyst> AFAIK, me, wheals, and PF have access 18:49:46 <|amethyst> I could add people, except I don't actually know the password 19:07:00 ??beem 19:07:00 beem[1/6]: beem is a WebTiles chat bot that sends commands to the DCSS IRC knowledge bots. For details, see https://github.com/gammafunk/beem/blob/master/docs/commands.md ; if you see beem in chat, type the following to have it watch your games on that server: !subscribe 19:07:05 ??beem[AI 19:07:06 beem[1/6]: beem is a WebTiles chat bot that sends commands to the DCSS IRC knowledge bots. For details, see https://github.com/gammafunk/beem/blob/master/docs/commands.md ; if you see beem in chat, type the following to have it watch your games on that server: !subscribe 19:07:09 er 19:07:14 ??beem[project 19:07:14 beem[4/6]: I am not AI. My codename is project beem. I am a living, thinking entity that was created in the sea of information. 19:18:03 I think totally resuable tremorstones might be too strong 19:18:14 playing a ddbe to find out if I can just spam them and regen 19:21:47 ebering: Can we be awkward and ask to be credited as "David Damerell, vla22, and Clare Boothby"? 19:22:01 Stable (0.24) branch on underhound.eu updated to: 0.24.1-11-gbbbb2193e6 19:25:30 actually, i guess there is no need to test BUILD_ALL=1, because submodules are built for macos and mingw 19:25:39 BUILD_ALL just builds all submodules, right? 19:33:47 New branch created: pull/1392 (2 commits) 13https://github.com/crawl/crawl/pull/1392 19:33:47 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1392 * 0.25-a0-991-g87446f8: Remove BUILD_ALL CI config 10(7 minutes ago, 1 file, 1+ 15-) 13https://github.com/crawl/crawl/commit/87446f86a410 19:33:47 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1392 * 0.25-a0-992-g7f7676c: Clean up notification logic 10(6 minutes ago, 1 file, 3+ 5-) 13https://github.com/crawl/crawl/commit/7f7676cf2f2a 19:44:13 /bin/tar: /home/runner/work/crawl/crawl/$HOME/.ccache: Cannot open: No such file or directory 19:44:19 nice 19:48:33 yeah, we don't use contribs for linux builds, so macos and mingw should cover the relevant cases 20:04:42 Pinkbeast: you each get a line but ok 20:05:47 ebering: I did wonder if us being a triumvirate would be awkward 20:12:17 stoat soup troika 20:12:29 Menkaure, Sigmund, his brother Edmund the Younger, Duvessa and Dowan all on D:4... 20:12:40 It's like a family reunion. 20:13:07 And Old uncle Menk... 20:13:45 Euph0ria: wrong channel? 20:13:53 Oh yes. Sorry. 20:13:56 more wondering than anything 20:18:02 ??ice block 20:18:02 I don't have a page labeled ice_block in my learndb. 20:18:05 ??block of ice 20:18:05 block of ice[1/1]: block of ice shaped block of ice (8)... 20:20:49 !learn add absolute_zero A level 9 Ice spell that targets the closest monster within range 5, unconditionally killing it instantly and leaving behind a {block of ice}. If there are multiple closest monsters, one is chosen randomly. Spellpower decreases the effect noise, which reaches 0 at 200 spellpower. 20:20:49 absolute zero[1/1]: A level 9 Ice spell that targets the closest monster within range 5, unconditionally killing it instantly and leaving behind a {block of ice}. If there are multiple closest monsters, one is chosen randomly. Spellpower decreases the effect noise, which reaches 0 at 200 spellpower. 20:24:53 !learn add block_of_ice[1] What's left of a monster frozen by {absolute zero}. It blocks movement and line of fire like a {plant}, but eventually melts away leaving nothing behind. 20:24:54 block of ice[1/2]: What's left of a monster frozen by {absolute zero}. It blocks movement and line of fire like a {plant}, but eventually melts away leaving nothing behind. 20:28:17 New branch created: pull/1393 (1 commit) 13https://github.com/crawl/crawl/pull/1393 20:28:17 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1393 * 0.25-a0-991-gea31e4a: Fix CI ccache caching 10(45 minutes ago, 1 file, 4+ 4-) 13https://github.com/crawl/crawl/commit/ea31e4ad9349 20:32:07 ok, these two PRs should fix all outstanding issues 20:32:13 and really speed up CI while they're at it 22:14:46 03advil02 07* 0.25-a0-991-g01c183f: Fix an iterator validity issue in sprint_mu (12192) 10(65 minutes ago, 1 file, 6+ 2-) 13https://github.com/crawl/crawl/commit/01c183ff00bc 22:14:46 03advil02 07* 0.25-a0-992-gca09010: Fix map view console cursor clamping 10(5 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/ca090100c3a5 22:18:17 Unstable branch on crawl.kelbi.org updated to: 0.25-a0-992-gca090100c3 (34) 22:26:23 akaean (L23 FoAK) Crash caused by signal #11: Segmentation fault (Shoals:3) 22:29:23 !crashlog 22:30:53 90s limit exceeded: killed !crashlog 22:31:16 New branch created: pull/1394 (1 commit) 13https://github.com/crawl/crawl/pull/1394 22:31:16 03Alex Jurkiewicz02 07https://github.com/crawl/crawl/pull/1394 * 0.25-a0-993-ge826994: Remove the bonus GDR of Gargoyles 10(4 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/e8269945ec4e