02:46:22 -!- Jetnerd_ is now known as Jetnerd 02:55:23 -!- amalloy is now known as amalloy_ 02:56:53 -!- amalloy_ is now known as amalloy 05:51:21 -!- amalloy is now known as amalloy_ 06:02:44 Hey, just wanted to say that I'm not 100% dead, or that I am upset with anyone in particular. The WJC dev proces took a significant chunk of my free time last year and I had to remove myself from it to pursue other things. I can't deny I'm upset about the direction (or lack thereof) it has grown, but I don't blame anyone in particular or hold any grudges. I do think the volunteer DCSS development process could do with more s 06:03:39 All that said, and knowing that I probably will be asleep when all of you are discussing, as per usual, I just want to say that I've been lurking around the later WJC discussions, and had some time to reflect on the bad parts of WJC myself, and I wrote a thing with some possible changes that could align with what the dev team wants 06:03:40 I may be completely off the mark, but at least I can try 06:03:42 https://pastebin.com/7piyxwNK 06:07:01 I focused my efforts here in making sure that Whirlwind and Wall Jump are both impactful and infrequent, by actively discouraging multiple uses in a fight, as well as discouraging their use in every fight 06:07:14 while keeping them simple interface wise 06:09:15 From what I've read from the dev team, these two issues (Whirlwind and Wall jump oscillating between useless and tediously mandatory) are the central problem with WJC's current design 06:09:51 so hopefully giving each a very particular, situational role that does not rely on cooldowns or usage restrictions solves that problem 06:12:09 And just a clarification for the "non-stacking" part of wall jump: what I mean is that you cannot increase the EV bonus by continuously jumping, it will only increase if you hit a higher number of enemies at once 07:20:27 better to use this paste as I can edit it https://pastebin.com/8YxKPhwh 08:26:15 SteelNeuron: I haven't read what you posted, but some days back there was a discussion about how to god might be retained, which I've summarized here: https://github.com/crawl/crawl/wiki/0.21-Development-and-Release-Plan#wu-jian-council-wjc-changes 08:26:46 yep I've caught up with that, that post is what I tailored my proposal towards 08:26:48 this was something |amethyst wanted to work on, so he might be a good person to coordinate with, but elliptic was also involved in the discussion 08:27:09 alright, I can take a look later then 08:27:14 cool :) 08:57:00 03Aidan Holm02 {advil} 07* 0.21-a0-427-gc5fe4b4: Fix sidebar for large tile_font_stat_size 10(12 days ago, 1 file, 10+ 23-) 13https://github.com/crawl/crawl/commit/c5fe4b46472e 09:08:30 03Aidan Holm02 {advil} 07* 0.21-a0-428-g08060ca: Show branch tiles in lookup-help menu 10(2 weeks ago, 3 files, 94+ 2-) 13https://github.com/crawl/crawl/commit/08060ca9ca5f 09:08:30 03Aidan Holm02 {advil} 07* 0.21-a0-429-gf361398: Add item tiles to lookup-help menu 10(2 weeks ago, 1 file, 22+ 2-) 13https://github.com/crawl/crawl/commit/f361398006bb 09:08:30 03Aidan Holm02 {advil} 07* 0.21-a0-430-g517d9b7: Add new card tile to lookup-help menu 10(2 weeks ago, 3 files, 17+ 2-) 13https://github.com/crawl/crawl/commit/517d9b72004a 09:09:43 merge storm PogChamp 09:20:43 aidanh: while testing that last one, I noticed that the messages window seems to be extending a bit to the right into the stuff in the lower right (I can reproduce this by doing ?/ in a game), not sure if this is related to some of your changes but it seems like potentially it's related to the clipping bug with the dungeon window that you fixed -- maybe the message window right edge is aligned with the dungeon? 09:37:51 03Yermak02 {advil} 07* 0.21-a0-431-g01fcc59: Fix identification of floor items in edge cases 10(6 days ago, 2 files, 8+ 0-) 13https://github.com/crawl/crawl/commit/01fcc5944612 09:38:31 The build was canceled. (master - 517d9b7 #8902 : Aidan Holm): https://travis-ci.org/crawl/crawl/builds/298564243 10:57:18 SteelNeuron: your first message got cut off, "I do think the volunteer DCSS development process could do with more st" 10:57:42 huh weird, I see it full 10:57:45 "structuring and better expectations however" 10:59:04 I think some clients don't fully enforce the irc char limit 11:09:13 we can't really enforce structure nor expectations 11:09:38 those things take a commitment from everyone that we just can't guarantee; that's the fundamental problem with volunteer dev 11:13:22 we could probably have better documentation of what to expect when submitting a PR 11:13:38 yeah 11:13:56 they would probably be somewhat similar to what we have in our new dev orientation 11:14:34 It's more what exactly Trunk means 11:14:56 I expected it to be a more flexible testing ground, while I found that once the god made it to Trunk, the resistance to testing changes was much higher than I expected 11:15:11 had I know that I would have pushed to keep it in an experimental branch for longer, but it barely got any dev attention there 11:16:08 right, there is the thing where, once it goes into trunk, it starts to become closer to being in a release; I think the difficulties arise when there are some fundamental problems with a feature from the standpoint of a number of devs 11:16:43 So that has to get resolved somehow (said devs agree on a compromise, or they or the original author make changes) 11:16:58 Sure, at that point however it's really hard for an outsider to have their changes tested at all 11:17:21 you see, I was not opposed to the god going in a different direction in Trunk, but what ended up happening is that it went in no direction at all 11:17:34 trunk feels a bit overprotected 11:17:39 just an outsider's perspective 11:17:49 Yeah it's important that the changes are moving things closer to a point where enough people are happy 11:18:17 if that doesn't happen quickly there can be problems 11:18:39 but like I said, it's just really hard to streamline a process, since people involved have free time...and then they don't 11:19:11 MPA wanted to work on the god some more, but didn't have time; elliptic (and I, and MPA) felt the god's design was not good enough for a release, and MPA didn't have time to fix 11:19:27 so the removal talk happened, but |amethyst wasn't happy with this, since he liked some of the design 11:19:45 so what it took was for elliptic to have the time to communicate his concerns 11:20:39 so before that happened he suggested removal, which I agreed to (along with moving some WJC things elsewhere) 11:20:59 but since |amethyst wanted to work on things, that meant we needed some compromise 11:21:18 so I'm sure about |amethyst's availability in the upcoming week for implementing stuff 11:23:20 I guess I need to read over this pastebin pretty carefully and chat with him and elliptic (and MPA if he's around) 11:23:58 but I would like to get this god somewhere reasonable so we can make it part of the release 11:26:29 I actually don't quite understand some of this new whirlwind description 11:26:39 Whirlwind pins the enemy in place for one "turn" (measured in your movement speed) after the enemy gets to act. 11:28:29 I'm not sure what status is happening to the monster here 11:29:14 it sounds as if it is applying a paralysis-like duration based on the time of the player's turn 11:29:41 ??whirlwind 11:29:42 wu jian[3/8]: Whirlwind: Triggered by moving from a tile adjacent to an enemy to another tile adjacent to it. 11:29:47 also in this: Whirlwind pins the enemy in place for one "turn" (measured in your movement speed) after the enemy gets to act. 11:29:55 er 11:30:03 in this: - Whirlwind does slightly lower damage than a normal attack (80%), with number of attacks compensated by attack speed. 11:30:14 number of attacks compensated by attack speed? 11:37:18 I don't have time to look at the proposed changes now (probably not for a few days at least), but I do think that testing things in trunk works a lot better when there is a dev or devs actually taking an active role with it 11:38:59 sorry, I'm only half online atm but 11:39:05 I don't mean paralysis, but rather inability to move 11:39:08 they can still attack 11:39:30 after the monster has their turn, they get 10 auts (or your speed equivalent) of a state that prevents movement 11:39:54 alright 11:40:21 (I don't know what the situation was when WJC was first put in trunk because I wasn't very active then, but I'm generally opposed to things being pushed to trunk just for testing if no dev is seriously going to work on it there) 11:41:21 yes, it sort of got merged but as a large feature didn't really have anyone taking ownership of it 11:42:56 So, for that short-term duration, it's going to be very short, so there will be instances of monsters of breaking out of it on the turn the player applies it 11:43:15 this would be the case for fast monsters that have higher than usual energy, I believe 11:44:11 which is not necessarilly a problem but it may cause some confusion 11:44:55 it's not a status like confusion where the monster's behaviour becomes unpredictable, so that's probably an improvement for something that's trying to reward careful movement choice 11:46:15 SteelNeuron: can you explain this phrase "with number of attacks compensated by attack speed." that appears for whirlwind and lunge? 11:47:08 That means no modification at all from the current system, which makes it so you attack an averaged number of times based on the ration between your move and attack speed 11:47:19 as of right now, if you move at 10 auts but attack at 5 auts, you'll attack twice per move 11:47:32 or 50% of the time if you move at 10 auts but attack at 20 auts 11:48:23 right, ok 11:53:05 maybe that status could be predicated on the monster's energy rather than your own? 11:53:11 not sure how that'd look, I've been out of the codebase for a bit 11:56:20 my intent is "they cannot move the turn after you used whirlwind" but I'm not sure that makes sense in the back end 11:58:55 right, for it to be sane you basically have to apply a duration with this status 11:59:18 I suppose it would just make them think that they're a stationary monster, which should be pretty reasonable to implement 12:00:12 but the duration needs to make sense; you can simply apply a fixed duration (possibly with some weak scaling based on something like piety, but since it's meant to be short, maybe any scaling is a bad idea) 12:00:51 I'm not sure if applying a duration based on player time or mosnter energy would work very well, I need to remind myself how monster enchants decay 12:01:19 I believe they decay according to the monster's energy and I think this is even independent of the time the player's turn took 12:09:40 Unstable branch on crawl.akrasiac.org updated to: 0.21-a0-431-g01fcc59 (34) 13:52:37 -!- amalloy_ is now known as amalloy 15:46:08 03gammafunk02 07* 0.21-a0-432-g6a4a694: Update descriptions of the Ogre species (Pereza0) 10(37 minutes ago, 2 files, 8+ 8-) 13https://github.com/crawl/crawl/commit/6a4a694a3ee1 17:14:37 arthad (L1 GrFi) ASSERT(!crawl_state.prev_cmd_keys.empty()) in 'main.cc' at line 3531 failed. (D:1) 17:14:48 !crashlog 17:15:01 18063. arthad, XL1 GrFi, T:0 (milestone): http://crawl.berotato.org/crawl/morgue/arthad/crash-arthad-20171107-221428.txt 17:19:15 what even is the "ins" key 17:19:49 advil: Insert, one imagines 17:20:24 !source main.cc:3531 17:20:24 https://github.com/crawl/crawl/blob/master/crawl-ref/source/main.cc#L3531 17:20:36 apparently it it 0 on a mac 17:20:40 *it is 17:20:58 oh my 17:21:20 <|amethyst> numpad ins and 0 are on the same key, toggled by numlock 17:21:48 oh. i thought you meant it was keycode 0 17:22:08 no, numpad 0 17:22:22 luckily (?) I also don't have numlock 17:22:49 but mac keyboards are normal usb keyboards, which means they do the backward compatibility dance anyway 17:23:14 (meaning it is likely to send a num lock keycode, which you may see depending on how raw raw mode is) 17:26:10 well, I got it to crash (as well as all sorts of other weird things) by doing repeats in a semi-random order on a new game, but I'm not entirely sure how 17:26:20 one time it saved, one time it put ?s over nearby items 17:28:34 oh maybe that last one is a gnoll thing 17:28:42 I guess they still have that ability 17:30:37 advil: ?s? i thought the mac terminal was better at unicode than that 17:30:51 oh. it just occurred to me, maybe you pasted in the right character and my irc client messed it up 17:30:53 I have tiles built now :) 17:31:02 it was just gn strong nose showing items 17:31:32 today's indecipherable emoji: ?s? 17:31:35 I've also now gotten it to repeat the repeat command once, but I'm not sure how I did that either 17:35:08 maybe that wasn't a clear response, strong nose actually uses tiles with the question mark symbol on item locations in tiles 17:39:07 i figured 17:40:18 whew, finally replicated that minimally...on a fresh game load, 0`` crashes the game 17:41:17 I challenge anyone to crash the game in fewer keystrokes 17:42:01 advil (L9 HOFE) ASSERT(!crawl_state.prev_cmd_keys.empty()) in 'main.cc' at line 3531 failed. (D:6) 17:42:21 and it indeed works in webtiles 17:44:52 at least it's not exploitable, I don't think it can happen without a save and a reload (or new game) 17:48:54 well, the second key can be any non-numeric 17:52:33 I think there might be something funky in general with the key replay stuff. I managed to temporarily hang a local game the other day by hitting some sequence (don't know what, wrong window had focus and crawl doesn't have view-lossage) 17:53:22 couldn't interrupt it, eventually it reported an error and went back to normal; I have a nasty suspicion it was trying to replay random memory and might potentially have crashed if it hadn't instead hit a sequence that triggered an error abort 17:53:28 er, playback abort 17:53:33 yeah, it seems that way, I'm not sure all the weird stuff I saw can be explained by the minimal case 17:54:23 since the bug I just found only happens when the key history is empty 17:56:11 i think there is funky stuff with key replay in general. i often hit ` several times, and on like the 10th time in a row it's like "what, you don't have any previous command to repeat". i'm pretty sure this is not a case of last turn's action becoming illegal 17:56:38 advil: &^C is two keystrokes, although of course this is cheating 17:57:03 3 keypresses 17:57:07 so I guess it's a tie 17:57:16 mine works outside of wizmode though 18:35:29 well, this is going pretty much how it goes every time I try to fix something in a part of crawl that is new to me 19:00:35 advil: I can't reproduce anything like that on the latest master 19:01:23 did you play with window size at all? 19:03:44 aidanh: here's a partial screenshot, https://www.dropbox.com/s/fc69aibjklovu1v/Screenshot%202017-11-07%2019.03.21.png?dl=0 19:04:01 fwiw this is on my retina computer, I haven't tried it on anything else 19:05:09 Yep, resizing caused no issues 19:06:02 Hmm, might be a retina issue then, although I'm not sure why that'd be the case 19:06:18 yeah, didn't seem likely 19:06:28 I can test it on a non-retina mac later 19:07:04 dcss font rendering does use high-dpi calculations though 19:07:43 so on your machine the linebreak gets inserted properly for long lines like that? 19:10:00 High-dpi calculations? I was under the impression dcss uses a font atlas for rendering 19:10:21 Yep, it wraps correctly; different text sizes are working fine here as well 19:11:33 %git da8ff79318f9b6e55ad1ac6304ce32a876208b18 19:11:33 07Grunt02 * 0.16-a0-2024-gda8ff79: Where high DPI allows, render fonts in high resolution. 10(3 years ago, 8 files, 42+ 21-) 13https://github.com/crawl/crawl/commit/da8ff79318f9 19:12:18 Hm, I'd have thought that just doubling the size of the font atlas would make high dpi work 19:12:45 oh, I see, that's pretty much what it does 19:13:01 yeah, just need to use device pixels instead of logical pixels 19:13:42 well the bug is probably the rest of crawl assuming that char_width() returns logical pixels 19:14:09 it doesn't seem off by the right multiplier to be that 19:14:37 though I'm not sure what else it would be, if it is a dpi issue 19:15:44 ... why do char_width + char_height return device pixels, btw? 19:16:34 so they do 19:17:13 maybe that's contributing...though the messages window is only ever off by 1cm or so (whatever that is in pixels on this screen) 19:18:18 that's why I was wondering if it was clipping related, because it looks like it could be breaking about where the dungeon ends if it weren't being clipped 19:19:44 I wonder if there's a way to simulate high-dpi mode on a non-high-dpi machine 19:19:46 ahh; that could be it 19:20:10 maybe that sets a max? 19:20:21 comment out the SDL GetDrawableSize call and substitute dummy variables 19:20:45 you're right that char_width being in device pixels ought to be causing some strange stuff 19:21:48 although only for people with retina screens, so it's a sneaky bug 19:22:16 doesn't get used too much it looks like 19:23:54 yeah, it's a retina issue; i can reproduce the clipping now 19:28:35 oh by your technique above? 19:30:25 I set *densityNum to twice the height of the window, instead of whatever the SDL call returned 19:32:25 ah nice, that's handy 19:32:49 Another thing I notice is that the WIZARD indicator is cut off on the right 19:33:28 from git grep it looks like a bunch of stuff in the hud uses char_width 19:37:31 the ratio's not 1/2 though 19:37:44 yeah, that's why I was puzzled 19:38:12 here is a theory: the width of the message window is calculated by text width, with a max of the width of the dungeon area 19:38:17 do you get the puple wiz indicator clipping on retina? 19:38:29 yep 19:39:24 comparing a sim-retina crawl with a normal version, i'm also seeing the sim-retina's fonts are rendered narrower on my screen 19:39:48 although not sure to what extent that's caused by an imperfect hack 19:40:40 they don't look overly narrow but it's hard to compare 19:41:20 *WIZARD* cuts off in the middle of the R on my retina display 19:42:23 that screenshot above was on retina so I guess you could compare with that 19:43:51 on second thought, that's probably an artefact of font rendering; if i fiddle with the font size it goes away 19:48:58 i'm also getting --more-- cut off by the bottom of the screen 19:50:06 oh, is that a retina thing? 19:51:21 I've noticed that 19:51:36 yeah, only on sim-retina 19:53:24 it's probably a bug in the font-size / text-size calculations; maybe something like using ints instead of floats for scaling 19:56:30 it's doing scaling with integer math by storing a fraction 19:56:36 but the ratio on a mac retina is 2 19:56:47 it's only on weird devices that you get a non-integer ratio 20:05:08 hmm, yeah 20:10:49 maybe it's something like font widths being scaled, but not advances 20:12:05 but it seems like all font rendering is done with logical pixels that are scaled on the gpu with a transform 21:12:13 03advil02 07* 0.21-a0-433-ge7131c6: Several fixes for repeat/again 10(12 minutes ago, 5 files, 52+ 15-) 13https://github.com/crawl/crawl/commit/e7131c60669a 21:12:13 03advil02 07* 0.21-a0-434-g806c6c2: Checkwhite / unbrace 10(9 minutes ago, 4 files, 6+ 7-) 13https://github.com/crawl/crawl/commit/806c6c2d46fd 23:56:30 any news on cjr console?