00:01:11 @gammafunk - it was me who talked about that, but I didn't get around to actually doing it. instead, I just tested on a branch on my fork :D 00:01:33 It's reasonably fast to test if you comment out all the other stages so that the one you want gets scheduled quicker :D 00:01:54 As for the brittleness of more CI checks - in this case I don't think so 00:02:16 Because the github action in question provides a docker container under the hood 00:02:29 if ldoc's dependencies change, they update their container image, and it just works 00:02:50 Whereas if we installed it into the GH VM ourselves, we'd be responsible for keeping up on that 00:03:07 But, tradeoffs, it's one more layer ofa bstraction 00:03:16 I tried the ldoc suggested way of installing from scratch first 00:03:23 it didn't work with several iterations of fiddling 00:03:31 I tried the nice GH action they supplied, it worked like second try :D 00:08:42 And, the nice thing about CI is that you do notice 00:09:17 Whereas if make api breaks, nobody notices until someone goes to update the docs by hand and you have to go figure out what changes over the past n weeks broke things :D 00:10:51 My worldview is that the overall rate of change of software mostly is constant. Sure, sometimes if you had waited to update someone else would have reported an issue and gotten it fixed first. But usually not. So it's just whether you're taking small changes frequently or big changes once in a while 00:11:04 small changes frequently seems to be easier to manage in my experience :D 00:13:03 Unstable branch on cbro.berotato.org updated to: 0.33-a0-788-gb9f7a0f2ca (34) 00:19:00 Anyway, I'm rambling. Glad to hear you think it's cool. It's not _just_ the lua docs though - I also publish the FAQ.html from make docs :) 00:54:32 Monster database of master branch on crawl.develz.org updated to: 0.33-a0-788-gb9f7a0f2ca 01:48:45 <04d​racoomega> So, I have been looking into giving confusion a proper icon (rather than looking the same as 'distracted', which ironically tends to confuse players) - in a similar fashion to what was done for paralysis a little while ago. And boy, the rat's nest I am finding when I look into this. Like, how mons_looks_distracted() is actually just 'monster is low-tier stabbable' and thus includes confused, blind, the player being invisible, etc. 01:48:45 And MB_DISTRACTED isn't the flag used to say a monster is distracted. That's MB_DISTRACTED_ONLY. MB_DISTRACTED is used to used show confusion on enemies (but MB_CONFUSED is actually used for friendly monsters who are confused). Also, several of these are used for hints mode in questionable ways. For instance, if you have stab highlights turned off, even hints mode won't tell you that a monster is asleep. Although the blurb lies anyway, and you can 01:48:45 trigger the following by paralyzing a monster after it's shouted in your face: Apparently it has not noticed you - yet. Note that you do not have to engage every monster you meet. Sometimes, discretion is the better part of valour. 01:49:13 <04d​racoomega> (And merely being asleep or friendly will trigger the `That monster looks a bit unusual. You might wish to examine it a bit more closely' hint >.>) 01:58:10 <04d​racoomega> (Also, the code for this sure is something: C++ bool mons_looks_distracted(const monster& m) { const stab_type st = find_stab_type(&you, m, false); return !m.friendly() && st != STAB_NO_STAB && !mons_looks_stabbable(m); } We want a monster that can be stabbed but also not. 01:58:23 Of course 01:58:29 <04d​racoomega> (Because mons_looks_stabble() is only for high-tier stabs) 01:58:37 STAB_NO_STAB is a pretty great enum member name tho 01:58:40 you have to give it that 01:59:53 Feels like there should a stab_quality enum 02:00:08 stab_quality::None|Low|High 02:00:13 instead of the denom thingy 02:00:20 then you get rid of mons_looks_stabble 02:01:30 <04d​racoomega> I actually thought there was no longer any point to the stab_type enum, but apparently this is used for something: noting action counts in logs. ie: how many types of which type of stab did you get. 02:02:00 <04d​racoomega> Which is mildly relevant, I suppose 02:02:34 <04d​racoomega> (Though there's at least one issue with that also, which is that it notes stabs you get against blinded monsters as 'Stab: Invisible'. Which may be mechanically true of the result, but feels untrue for bookeeping about player actions) 02:03:05 Yeah, that feels like it should be separate, though it rarely comes up I suppose 02:03:10 how often do you blind a monster 02:03:27 <04d​racoomega> Potentially quite often if you're an enchanter! 02:03:48 Hmmm. Not the enchanter attempts I've made but that probably means i've been doing it wrong 02:03:53 all the guides are pretty old for En 02:04:27 <04d​racoomega> (I mean, confusing touch is more of your bread and butter against things it's good against, but dazzling flash very specifically hits a bunch of things that confusing touch will not. And at higher levels, the way Enfeeble enables stabs is via blind) 02:04:35 Oooh, right 02:04:42 forgot that did blinding :D 02:04:42 have used it 02:04:47 confusing touch is kinda a stab 02:04:51 a trap rather 02:05:50 Which maybe needs to be addressed somehow 02:05:59 Or maybe not, but it's kinda a noob trap 02:06:12 in that if MR is high, but not TOOO high 02:06:22 you waste n turns trying to confuse 02:06:49 <04d​racoomega> The fact that using confusing touch against monsters with very high will is actively bad for the player is something that has been discussed a few times. (But the spell is still quite strong for a while, despite this.) 02:07:03 Quite strong for a while, but then murderous 02:07:53 but I see your point 02:11:21 And it's somewhat a more general challenge with hexes isn't it? 02:11:27 high will I mean 02:11:33 Yes I know orb of guile 02:12:33 And yes Dazzling Flash doesn't check will 02:13:37 I've never gotten an En far enough for Discord tho 02:13:52 I suck at playing enough that I am not well suited to speaking to balancing :D 02:20:22 <04d​racoomega> find_stab_type() also has arguments for a non-player attacker (which can even be null) and a non-monster victim, even though neither can ever be true in practice 02:20:42 <04d​racoomega> (I don't recall if monsters ever could get non-sleep-stabs against anything) 02:20:54 <04d​racoomega> (Those don't use this codepath at all, either) 02:36:05 <03i​mplojin> yes i've run into that stab stuff a few times and it's just awful 02:37:30 <04d​racoomega> I am going to do my best to untangle at least some of it 02:37:33 <03i​mplojin> one possible hiccup to removing that stab enum is that iirc those stab states are the only thing we use to distinguish stab types in console builds? would have to look at it again 02:38:02 <04d​racoomega> Well, stab_type is mapped onto color flags used in console, but that can obviously be done any other way 02:38:35 <04d​racoomega> The only reason (I can see) to have more fine-grained stab_type than low-tier/high-tier/none is for action count tracking, but that is mildly interesting, at least 02:38:59 <04d​racoomega> (So rather than cut it, I was going to make it actually properly distinguish blind and invis stabs) 02:41:56 <04d​racoomega> Not entirely sure what to do about some hints mode messages. Like, a few of the current ones are sort of nonsense here, but not always clear what to replace them with. (It also seems buggy in some other ways, such as hilighting the monster it's talking about only after the monster moves away from the hilight??) 02:42:08 <04d​racoomega> Or rather, hilighting where it used to be standing, once it leaves 02:42:15 <04d​racoomega> Deeply confusing 02:42:44 <04d​racoomega> But I don't think it should claim that something about a monster looks unusual just because it's asleep. That's like one of the most usual ways to encounter a monster 02:44:30 <04d​racoomega> Like, That monster looks a bit unusual. You might wish to examine it a bit more closely by hovering your mouse over its tile. pressing $cmd[CMD_LOOK_AROUND] and moving the cursor onto its square. is seemingly printed for the first stabbable (or friendly) monster you see. 02:45:13 <04d​racoomega> Tempted to just hijack that for the unusual monster hilight, tbh >.> 02:45:17 <03i​mplojin> yes 02:45:32 <03i​mplojin> should change that wording now that unusual is a thing, for sure 02:45:46 <03i​mplojin> didn't realize we had that as a hints message, alas 02:46:24 <04d​racoomega> I mean, hints mode is mostly ancient 02:46:38 <04d​racoomega> (Not clear how many people use it, either, though its existence is largely positive as long as it's not too out of date) 02:46:49 <04d​racoomega> But some of the advice here seems very dubious in other ways 02:47:26 <04d​racoomega> Like, actually distracted monsters can get: ostr << "Apparently it has been distracted by something. You could " "use this opportunity to sneak up on this monster - or to " "sneak away."; which feels factually untrue about how that works. 02:47:57 <04d​racoomega> Implying things about Crawl's stealth that just don't really reflect reality 02:49:38 <03i​mplojin> also, +1 to doing clua docs in ci, hopefully it will help folks to notice if a commit breaks ldoc 02:51:57 <04d​racoomega> Some of this probably predates the tutorial, doesn't it? (Surely we at least tell people to xv things there, no?) 02:52:41 <03i​mplojin> pretty sure we do, i have a branch somewhere where i was starting on a revision pass for the tutorial levels 02:53:23 <03i​mplojin> i'd taken hints mode as being intended as a kind of tutorial-lite, to be honest 02:56:04 <04d​racoomega> I mean, I am not saying that it doesn't bear reminding here 02:56:17 <04d​racoomega> (Player could definitely bear being told to examine things more often :P) 02:56:40 <04d​racoomega> But I don't think "Oh, that monster is asleep; better make sure!" is the best time to sound the warning bells, perhaps ^^; ) 02:57:08 <04d​racoomega> (I am not going to do much to hints mode as I clean this up, but will at least try to make a couple related things less obviously incorrect) 02:57:29 <04d​racoomega> Where the refactoring has to touch them anyway 02:57:52 <03i​mplojin> definitely could see replacing that bit with, like, a venom dagger or whatever's appropriate and introduce it as xv for unusual item instead 02:59:55 <04d​racoomega> Yeah, I am taking out the hint for seeing a 'stabbable' monster (which feels kind of worthless), putting one in for an unusual hilight monster, and lowering the necessary danger threshold for a monster being 'interesting' to yellow instead of red. 03:00:14 <04d​racoomega> Yellow is worth saying "You could check them for further details" 03:00:21 <03i​mplojin> mmmm 03:00:38 <03i​mplojin> is this paired with changing default threat highlight to display yellow? we don't do that by default iirc 03:00:50 <03i​mplojin> (maybe we should?) 03:04:23 <04d​racoomega> I wasn't planning on it. (Honestly, I think the way the tiers broaden so much at higher XL is already a bit weird. It's very, very hard to get things to be red at high XL, even when they're super dangerous, and stuff stays yellow approximately forever. A few times the topic has come up in past, I was very tempted to try to decouple it from XP value and be able to specify some arbitrary 'threat tier' in monster data, which could be 03:04:23 manipulated in a way that better corresponds to how dangerous they should be considered relative to their depth, if not for the fact that Oka and Sac Courage both mechanically depend on it very much >.>) 03:05:03 <04d​racoomega> It's slightly awkward that this sort of serves double-duty as a UI reminder and something with mechanical impact 03:05:52 <04d​racoomega> Because there are monsters that have a support role that is dangerous, but aren't really 'Duel-worthy' exactly 03:07:28 <04d​racoomega> (Sac Horror also has somewhat weird thresholding where a major penalty can turn into literally 0 with one XL difference) 03:07:33 <04d​racoomega> Er 03:07:34 <04d​racoomega> Sac Courage 03:07:58 <04d​racoomega> (But basically, I've had an entire screenful of enemies all stop counting at once, with one more XL) 03:09:01 <04d​racoomega> (When I removed stat drain, I experimented briefly with XXX having a chance of giving you the Horror mutation temporarily, but it felt too janky and inconsistent) 03:09:34 <04d​racoomega> Actually, I thought it literally wasn't working at first, but it apparently didn't consider that screenful of things worth being afraid of at all until I was one XL lower 😛 03:10:08 <03i​mplojin> ha 03:13:01 <03i​mplojin> /users 03:27:31 <03i​mplojin> gonna merge this tile overlay PR, seems probably fine, if for some reason the optimizations here don't work as intended on cao's ancient gcc we can revisit 03:31:46 03Cgettys02 {Implojin} 07* 0.33-a0-789-g599c4360dd: fix: overflow of overlays with Qazlal's Disaster area (prawnwizard) 10(6 days ago, 7 files, 152+ 92-) 13https://github.com/crawl/crawl/commit/599c4360dd1e 04:33:49 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-5261-gd9800d219b 05:10:17 <04d​racoomega> (I actually thought I'd cherry-picked a couple of these fixes days ago now >.>) 05:10:34 03DracoOmega02 07* 0.33-a0-790-g7f59f0ee2d: Update artprops when swapping between talismans of same type (Ogregutan) 10(3 days ago, 1 file, 6+ 0-) 13https://github.com/crawl/crawl/commit/7f59f0ee2d87 05:10:34 03DracoOmega02 07* 0.33-a0-791-gf901164d30: Fix !lig turning flight items off (Ge0FF) 10(2 days ago, 2 files, 4+ 3-) 13https://github.com/crawl/crawl/commit/f901164d3028 05:10:34 03DracoOmega02 07* 0.33-a0-792-g9dea29b9c6: Fix dancing weapons made by Tukima's Dance being abjurable (regret-index) 10(30 hours ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/9dea29b9c6d5 05:10:34 03DracoOmega02 07* 0.33-a0-793-g5d0df7669e: Fix wand charges not always displaying for held wands in local tiles 10(3 hours ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/5d0df7669e27 05:10:34 03DracoOmega02 07* 0.33-a0-794-g5f9c754f44: Give confusion its own icon, refactor some stab-tier related code 10(25 minutes ago, 19 files, 78+ 106-) 13https://github.com/crawl/crawl/commit/5f9c754f442e 05:21:52 Unstable branch on crawl.akrasiac.org updated to: 0.33-a0-789-g599c436 (34) 08:38:39 03DracoOmega02 07* 0.33-a0-795-gc6a1a9b5a5: Fix new confusion/unaware icons not rendering in webtiles (Kab) 10(3 minutes ago, 1 file, 2+ 1-) 13https://github.com/crawl/crawl/commit/c6a1a9b5a586 08:38:39 03DracoOmega02 07* 0.33-a0-796-g61249f3da4: Reduce redundant messages when examining wandering monsters 10(2 minutes ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/61249f3da454 08:38:39 03DracoOmega02 07* 0.33-a0-797-g2b3b4944bb: Simplify a clua stab tier function 10(80 seconds ago, 1 file, 2+ 18-) 13https://github.com/crawl/crawl/commit/2b3b4944bb95 14:00:08 <06w​ensley> considering giving lower D the Depths treatment and splitting it off into its own branch? I have some idle ideas 😛 14:29:23 -!- ssm_ is now known as ssm 14:41:12 Drazool (L24 DjFE) ERROR in 'beam.cc' at line 745: ASSERT failed: flavour of 0 out of range BEAM_NONE + 1 (1) .. BEAM_FIRST_PSEUDO (101) (Depths:2) 16:42:20 Unstable branch on underhound.eu updated to: 0.33-a0-797-g2b3b4944bb (34) 23:02:33 <06m​umra> It sort of seems like that's the opposite of what's being said (and would possibly detract from what makes Depths exciting) 23:05:57 <06m​umra> (I do have some Depths monsters kicking around as well ... when I saying "throughout D" i am still mentally including Depths in my definition of D) 23:35:22 Unstable branch on crawl.develz.org updated to: 0.33-a0-797-g2b3b4944bb (34) 23:50:04 My 2 cents, it could make sense 23:50:08 as a terrible player 23:50:21 the whole "oh clear thru D:11 then this then that" is always a bit weird 23:50:34 like there's pretty much 1 linear ordering everyone follows 23:50:39 but it's not actually linear 23:56:14 <06m​umra> hmm i see yeah 23:56:27 <06m​umra> what would you call it though ... Mids? 23:56:50 Depths, and make Depths The Dark? 23:56:53 for alliteration 23:58:21 More honestly though I have no idea :D 23:58:35 Windows builds of master branch on crawl.develz.org updated to: 0.33-a0-797-g2b3b4944bb 23:59:15 My creativity more lies in the raw coding department (refactoring w/o breaking anything, debugging, simplifying) than it ever has been at naming things or game design :D