00:00:03 <Eksell> Unstable branch on CRAWL.XTAHUA.COM updated to: 0.20-a0-1020-g7b9e4fe (34)
00:34:54 <Cheibriados> Win not registered on profile 13https://crawl.develz.org/mantis/view.php?id=11019 by micr0chasm
01:17:04 -!- yesno__ is now known as yesno
01:28:37 -!- grisha5 is now known as grisha5|away
01:46:47 -!- amalloy_ is now known as amalloy
01:49:13 -!- grisha5|away is now known as grisha5
01:52:15 -!- Bammboobies is now known as Bammboo
02:43:14 -!- debo_ is now known as debo
02:44:11 <Cheibriados> Game mouse cursor doesn't line up with OS mouse cursor 13https://crawl.develz.org/mantis/view.php?id=11020 by bed
02:50:07 -!- yednesdaw is now known as wednesday
03:02:12 -!- amalloy is now known as amalloy_
03:12:57 <Rotatell> Unstable branch on crawl.beRotato.org updated to: 0.20-a0-1020-g7b9e4fe (34)
04:03:38 <Cheibriados> 03gammafunk02 07* 0.20-a0-1021-gcba205c: Changelog through 0.20-a0-1020-g7b9e4fe 10(4 hours ago, 1 file, 24+ 7-) 13https://github.com/crawl/crawl/commit/cba205c31713
05:10:08 <Jorgrell> Unstable branch on crawl.jorgrun.rocks updated to: 0.20-a0-1021-gcba205c (34)
06:21:03 <Eksell> Gorglomux (L5 FeEE)  (D:3)
08:17:06 <Cheibriados> Yara's Violent Unravelling uses AoE targetting 13https://crawl.develz.org/mantis/view.php?id=11021 by Leszczynek
08:53:22 <Cheibriados> Random console disconnects, broken pipes. 13https://crawl.develz.org/mantis/view.php?id=11022 by Le_Nerd
08:55:46 <advil> I'm sure that'll be easy to replicate
09:00:34 <advil> though I doubt it's crawl-specific
09:03:24 -!- amalloy_ is now known as amalloy
09:05:45 <|amethyst> !crashlog * cdo
09:05:58 <Sequell> 2646. ceb, XL23 GrFi, T:46554 (milestone): http://crawl.develz.org/morgues/0.19/ceb/crash-ceb-20170319-024400.txt
09:45:36 <advil> is this mouse position thing new?
09:47:10 <advil> I've never noticed it before, and it's really bad
09:47:23 <|amethyst> mouse position?
09:47:36 <advil> !bug 11020
09:47:36 <Sequell> https://crawl.develz.org/mantis/view.php?id=11020
09:47:47 <advil> reported last night
09:49:25 <advil> I don't play a lot of webtiles but I have tested on it in the last few weeks and I feel like this would be hard to miss
09:49:41 <|amethyst> it probably happens only in some browsers
09:50:05 <advil> well, the reporter replicated it on a bunch, and I'm seeing it in firefox now
09:50:16 <|amethyst> last relevant changes were
09:50:16 <|amethyst> %git 3925e26
09:50:16 <Cheibriados> 07edlothiol02 * 0.16-a0-2708-g3925e26: Webtiles: Do device pixel scaling better, fixing Firefox rendering. 10(2 years, 5 months ago, 2 files, 20+ 25-) 13https://github.com/crawl/crawl/commit/3925e268bea4
09:50:16 <advil> which is what I normally use
09:50:17 <|amethyst> %git 0475afac
09:50:17 <Cheibriados> 07Grunt02 * 0.16-a0-2348-g0475afa: Webtiles: scale canvases by devicePixelRatio properly. 10(2 years, 5 months ago, 5 files, 37+ 17-) 13https://github.com/crawl/crawl/commit/0475afac5cde
09:50:23 <advil> hmm
09:50:38 <advil> maybe I just never use the mouse
09:51:11 <|amethyst> using the mouse is recent
09:51:14 <|amethyst> so I guess that's it
09:51:25 <|amethyst> also, I fixed the <pre> tags in that bug report
09:51:31 <advil> heh thanks
09:52:13 <|amethyst> yeah, on the rare occasion that I use webtiles, I don't generally use the mouse
09:52:32 <|amethyst> so I never would have noticed
09:52:58 <advil> I'm really pretty sure I would have noticed this when testing the berserk screen flashing on webtiles
09:53:11 <advil> because I was using the mouse a bunch there (it forces refresh on the tile it hovers over)
09:54:03 -!- amalloy is now known as amalloy_
09:56:14 <Cheibriados> 03advil02 07* 0.20-a0-1022-g99c12e3: Replace spell success breakpoints with a polynomial 10(7 days ago, 2 files, 61+ 30-) 13https://github.com/crawl/crawl/commit/99c12e3834af
09:56:14 <Cheibriados> 03advil02 07* 0.20-a0-1023-gc4d110f: Remove linearity breakpoint for raw fail rate 10(2 days ago, 1 file, 26+ 18-) 13https://github.com/crawl/crawl/commit/c4d110fe692e
09:57:36 <advil> that polynomial may need further tweaking, I think it could be a slight buff (1-2%) for spell fail rates in some cases
09:59:23 <|amethyst> advil:  what you might do
09:59:57 <|amethyst> advil:  plot what the old value was for every reasonable input value
10:00:13 <|amethyst> advil:  (minmay and I used just the places where it changes)
10:00:42 <|amethyst> advil:  and then curve-fit that.  The fit won't be as "good" because it's piecewise constant, but it will more closely match the actual data
10:01:05 <|amethyst> s/it's/the curve you're fitting against is/
10:01:06 <advil> I did a lot of spot-checking of that (except looking at final fail rates rather than raw)
10:01:33 <advil> but yeah, I haven't plotted the exact stepped function
10:02:48 <|amethyst> there's no real reason for that 262144 to be a power of two, this code doesn't run in a hot loop or anything
10:02:59 <|amethyst> so you could change it to 260000
10:03:25 <|amethyst> which should increase (raw) fail rates by a little less than 1%
10:03:31 <|amethyst> and tweak as necessary
10:04:33 <advil> I might also switch it back to the non-Horner's version, it's easier to tweak :-)
10:06:11 <advil> %git 330ea1e6936505e2208336d7e44f8ba98b6a3893
10:06:11 <Cheibriados> 07wheals02 * 0.19-a0-1936-g330ea1e: Make webtiles mouse control an option (off by default). 10(6 months ago, 4 files, 12+ 1-) 13https://github.com/crawl/crawl/commit/330ea1e69365
10:07:06 <advil> oh that's just the option
10:09:19 <Jafet> gcc probably rewrites it anyway
10:09:19 <Jorgrell> Unstable branch on crawl.jorgrun.rocks updated to: 0.20-a0-1023-gc4d110f (34)
10:18:12 <advil> hmm, I can't replicate the mouse issue locally, in the exact same browser that I can clearly see it on CBRO with no effort
10:21:06 <advil> ^rc advil
10:21:07 <Sequell> http://crawl.berotato.org/crawl/rcfiles/crawl-git/hyperadvil.rc
10:22:21 -!- amalloy_ is now known as amalloy
10:25:25 <amalloy> Jafet: i was thinking that myself, but i wouldn't be at all surprised to learn that there are reasons it can't, eg because the two versions might behave differently in the presence of integer overflow or underflow
10:27:10 <|amethyst> advil:  IMO they're just as easy to tweak, the coefficients are listed right there anyway
10:27:38 <advil> true enough, but, you have to understand horner's rule to believe that
10:27:39 <|amethyst> the numbers are the same either way I mean
10:27:46 <advil> aha, it's an interaction with browser zoom
10:27:51 <advil> the mouse cursor thing that is
10:28:34 <amalloy> advil: oh, i didn't realize that wasn't in the bug report. i would have mentioned it myself if i'd known you were trying to figure that out
10:28:56 <amalloy> yes, browser zoom fucks with like everything that tries to track the mouse with js
10:29:08 <advil> yeah, I'm not sure the reporter realized that
10:29:16 <amalloy> on the one hand it sucks that we get it wrong; on the other, we are in good company
10:29:33 <amalloy> advil: imo mark as Infeasible until google drive figures it out :P
10:29:41 <advil> heh
10:29:54 <advil> I'll just mark it as confirmed
10:46:56 <Jafet> amalloy: int overflow is UB, so compilers will rearrange code pretty indiscriminately (some examples here https://kristerw.blogspot.de/2016/02/how-undefined-signed-overflow-enables.html)
13:24:50 <Henzell> Unstable branch on crawl.akrasiac.org updated to: 0.20-a0-1023-gc4d110f (34)
13:54:19 <Cheibriados> Reloading after a crash causes items to duplicate 13https://crawl.develz.org/mantis/view.php?id=11023 by hawthorn
13:56:59 <wednesday> Does anyone know the server admin on crawl.akrasiac.org? The ssh seems to not work. I can't do anything once I login. W,L,R nothing works. Eventually the connection times out.
13:58:30 <geekosaur> johnstein, |amethyst ^
13:59:30 <wednesday> geekosaur: was that for me?
13:59:41 <geekosaur> on your behalf :)
14:00:14 <geekosaur> both of them have access to cao
14:00:56 <wednesday> they guys over in ##crawl said this might be the palce to get a hold of the admin. maybe I'm the only one with the problem or the admin is out.
14:01:16 <wednesday> geekosaur: thank you.
14:01:24 <geekosaur> well, not everyone is around at all times, I was just highlighting them so hopefully they'll notice
14:02:44 <wednesday> I understand. that's the point of a server. :) fire it up and let it do its thing.
14:04:39 <amalloy> i confirm cao ssh is unresponsive at this time (to new logins at least)
14:07:01 <wednesday> I can login but I just can't do anything once there. I would have registered again to get the admin info. I think registration shows who to contact but no response.
14:07:21 <geekosaur> right, it should point you to amethyst as the server admin
14:07:35 <wednesday> I thought maybe I would find Rax but I heard Rax hasn't been around for years.
14:08:06 <geekosaur> admins the hardware, around but we avoid pinging them
14:08:19 <geekosaur> amethyst and johnstein are the points of initial contact
14:08:48 <wednesday> geekosaur: thanks. I'll try to contact them here. Maybe they will answer.
14:12:37 <amalloy> wednesday: geekosaur's already sent up the signal flare to try to contact them by mentioning their usernames (and was asking that you not try for any more attention from r.ax by mentioning that username again). the thing to do now is wait: |amethyst usually at least checks messages every day. the other productive thing you could do is open a mantis ticket in case these ephemeral IRC messages miss their marks
14:13:24 <wednesday> I sent them a message maybe they will get it.
14:14:15 <wednesday> amalloy: thank you. Sorry I'm not that good with IRC.
14:15:00 <wednesday> and I'm pretty knew to the crawl community way of things, although I've been trying to understand the game for a long time.
14:15:18 <wednesday> I will wait.
15:38:59 -!- fazisi is now known as not_mibe
16:47:26 <|amethyst> wednesday:  hm, I have no problem connecting
16:47:40 <|amethyst> wednesday:  maybe whatever it was unbroke itself already?
16:50:24 <amalloy> |amethyst: still broken for me
16:50:28 <amalloy> i can connect, but keystrokes do nothing
16:50:46 <amalloy> hm
16:50:57 <geekosaur> that was the report, yes, could connect but keys didn't work
16:50:59 <amalloy> actually it's just (W) that's broken
16:51:04 <amalloy> i can log in, and i can play
16:51:06 <amalloy> but i can't watch
16:51:47 <|amethyst> ohh
16:52:07 <geekosaur> "W,L,R nothing works"
16:52:30 <|amethyst> probably tried W first, which freezes the whole connection
16:52:47 <|amethyst> which usually means something screwed up with dgl's shmem
16:53:27 <|amethyst> fixed
16:53:34 <geekosaur> wednesday, ^
16:54:06 <|amethyst> wednesday:  see if it works now.  The problem was that 'W' was broken, and froze up the whole connection.  That should be fixed.
17:03:52 <CanOfWorms> https://www.reddit.com/r/dcss/comments/65sqoi/bat_form_challenge_done/
17:03:54 <CanOfWorms> dang
17:04:04 <CanOfWorms> sadly not the true bat challenge
17:04:07 <CanOfWorms> ??bat challenge
17:04:07 <Sequell> bat challenge[1/2]: Die to a bat as a deep dwarf.
17:04:10 <CanOfWorms> ??bat challenge[2
17:04:11 <Sequell> bat challenge[2/2]: No, berserk doesn't work. And inner flame is cheating. Forms don't work either!
17:04:17 <CanOfWorms> uh, the other bat challenge.
17:06:54 <kogasa> get lethally poisoned by a redback, then draw degeneration until it turns into a bat, then die
17:07:22 <kogasa> i think redback poison can breach damage shaving anyway
17:07:51 <CanOfWorms> !lg * dd killer~~bat
17:07:52 <Sequell> 75. KevinV12000 the Digger (L6 DDEE of Sif Muna), hit from afar by a battlesphere (created by SeriaLain's ghost) in Ossuary (ossuary_tomb_2) on 2017-03-26 18:10:43, with 357 points after 3574 turns and 0:39:49.
17:08:03 <CanOfWorms> !lg * dd killer=bat
17:08:04 <Sequell> 4. Tridious the Covered (L1 DDFi), blasted by a bat (exploding inner flame) (hexed by the player character) on D:1 on 2016-03-15 01:18:12, with 7 points after 356 turns and 0:04:01.
17:08:30 <CanOfWorms> !lg* dd killer~~"bat shaped"
17:08:36 <CanOfWorms> !lg * dd killer~~"bat shaped"
17:08:37 <Sequell> No games for * (dd killer~~'bat shaped').
17:08:44 <CanOfWorms> hmm
17:26:59 <wednesday> Its fixed! :) crawl.akrasiac.org ssh interface is working again.
17:37:57 <|amethyst> yay
17:44:37 <wednesday> Thank you all!
18:15:41 <Lantell> Unstable branch on underhound.eu updated to: 0.20-a0-1023-gc4d110f (34)
18:26:27 -!- Dracunos_ is now known as Dracunos
20:19:48 -!- amalloy is now known as amalloy_
20:46:51 <Cheibriados> 03advil02 07* 0.20-a0-1024-ge4c44be: Check exclusions for wall jump at landing spot, not at the wall 10(6 minutes ago, 1 file, 10+ 2-) 13https://github.com/crawl/crawl/commit/e4c44be3fceb
21:02:51 <ProzacElf> c - tha +5 falchion "Widowermaker" (weapon) {speed, +Blink rN+ Int-4}
21:02:55 <ProzacElf> excellent name
21:09:22 <Jorgrell> Unstable branch on crawl.jorgrun.rocks updated to: 0.20-a0-1024-ge4c44be (34)
21:41:27 <advil> !source player_spec_death
21:41:27 <Sequell> 1/1. https://github.com/crawl/crawl/blob/master/crawl-ref/source/player.cc#L1679
21:41:42 <alexjurkiewicz> beat you :)
21:41:53 <advil> heh
21:46:31 <wheals> left a note on #11020
21:47:13 <advil> oh thanks wheals, you are who I was figuring would most likely know something
21:47:24 <wheals> unfortunately, i don't know very much :)
21:47:35 <advil> heh
21:47:37 <wheals> the code was originally written by edlothiol i think? but they're gone
21:47:58 <advil> browsers are very mysterious
21:49:02 <wheals> grunt did some DPI stuff too
21:49:26 <wheals> though i figure you found it via git blame already
21:49:30 <wheals> %git 75d9865c86980b792806e6248855cba6e09b9c1a
21:49:30 <Cheibriados> 07Grunt02 * 0.16-a0-2106-g75d9865: Webtiles: scale dungeon canvas by devicePixelRatio. 10(2 years, 5 months ago, 2 files, 6+ 2-) 13https://github.com/crawl/crawl/commit/75d9865c8698
21:50:34 <wheals> wait, that got reverted
21:50:36 <wheals> %git 7e03aa0adb56bf9efec91f7d9e2b037749f86c48
21:50:36 <Cheibriados> 07Grunt02 * 0.16-a0-2156-g7e03aa0: Revert "Webtiles: scale dungeon canvas by devicePixelRatio." (#9189) 10(2 years, 5 months ago, 2 files, 2+ 6-) 13https://github.com/crawl/crawl/commit/7e03aa0adb56
21:50:44 <wheals> this was the correct one
21:50:45 <wheals> %git 3925e268bea408accdad6c7c53f37f2d36b3116a
21:50:45 <Cheibriados> 07edlothiol02 * 0.16-a0-2708-g3925e26: Webtiles: Do device pixel scaling better, fixing Firefox rendering. 10(2 years, 5 months ago, 2 files, 20+ 25-) 13https://github.com/crawl/crawl/commit/3925e268bea4
21:51:24 <advil> I hadn't found that
21:52:34 <advil> hmm, this stuff is all black magic
21:52:51 <wheals> maybe scaling the cursor position by devicePixelRatio would help?
21:53:04 <wheals> this is already farther than i got when i first ran into the bug
21:54:25 <advil> yeah, that seems like a plausible idea.  I'm worried that anything that makes it work on ff/chrome will break safari.
21:55:18 -!- MadCoyote is now known as FunkyBomb
21:55:21 <advil> though maybe safari's zoom just doesn't change the devicePixelRatio at all
21:55:46 <wheals> looking at MDN docs, it doesn't say whether mouse coords are in physical pixels or CSS pixels, which could mean it's unspecified...
21:56:54 <advil> nice
21:57:18 <advil> amalloy pointed out that google docs of all things had effectively this bug for years (I couldn't tell if it still does)
21:57:25 <advil> so it might not be solvable
21:57:55 <wheals> i use google docs on permanent 80% zoom and have had no problems, but for all i know that only works on firefox
21:58:20 <advil> I didn't find any reports in the last few years so I suspect it did get fixed or worked around
21:58:28 <advil> couldn't figure out how, unfortunately
21:59:40 <wheals> scaling devicePixelRatio by cursor is a lead, i guess; unfortunately, i don't have a linux setup easy accessible atm so i can't test webtiles really
21:59:59 <wheals> if it does get fixed, it'd be nice to have the webtiles mouse in stable
22:00:22 <advil> I bet this will interact with retina displays too
22:00:26 <wheals> i disabled it outside of trunk because the bug seemed too severe and i didn't know how to fix it, and there was a release not long after
22:00:50 <advil> ah I see
22:01:08 <advil> I can try that and test webtiles at some point
22:01:30 <advil> wonder if any of us have a retina display
22:04:32 <wheals> it's funny, i only saw the bug report because mantis has a bug where issues being "updated" sends you an email even if you have most stuff disabled
22:04:34 <wheals> serendiptious
22:24:45 -!- sesh is now known as Guest85527
22:24:52 -!- Guest85527 is now known as seshions
22:33:18 <advil> https://bugzilla.mozilla.org/show_bug.cgi?id=809788
22:33:26 <advil> "RESOLVED WONTFIX"
22:33:43 <advil> "it sounds like we should stick with our (and IE9's) existing behavior, and expect/hope that webkit and opera will bring their implementations into line."
22:33:49 <advil> thanks mozilla
23:09:02 <alexjurkiewicz> advil: i had retina, what do you want to test?
23:09:52 <advil> one question is, do you have weird issues with mouse cursor position in webtiles?
23:10:01 <alexjurkiewicz> not in chrome
23:10:08 <alexjurkiewicz> actually, let me test
23:10:28 <advil> also, I'm going to push a webtiles fix shortly that will need tested on retina, sooner or later
23:10:42 <alexjurkiewicz> yes, it's offset about -8,-5
23:10:45 <advil> specifically in trunk, w.r.t. selecting tiles
23:10:46 <advil> ok
23:11:22 <advil> this fix will hopefully address retina too
23:12:31 <advil> thanks for checking
23:18:31 <Cheibriados> 03advil02 07* 0.20-a0-1025-g4706f7e: Scale cell widths/heights by devicePixelRatio (wheals) 10(17 minutes ago, 1 file, 6+ 2-) 13https://github.com/crawl/crawl/commit/4706f7efea59
23:19:51 <advil> wheals: I think that worked perfectly (technically I scaled the cell dimensions but it's effectively scaling the mouse position)
23:20:06 <wheals> nice
23:20:33 <alexjurkiewicz> trog doesn't give a warning when you memorise a spell before you enter penance
23:21:09 <wheals> oh wow, 0.16 was 2014
23:21:18 <wheals> how time flies
23:22:28 <advil> heh
23:23:02 <advil> I think 0.16 is the version I have the most wins with
23:23:15 <advil> it's been downhill since then
23:23:35 <wheals> sounds like #11020 can be closed, though the code quality complaint remains unfixed
23:23:49 <advil> well, that's why I added that link to sgrunt's commits
23:23:57 <advil> because it does seem like we do that scaling for a good reason
23:24:02 <advil> or a reason
23:24:46 <advil> perhaps there's a better way, but I'm ok with this one if it worked
23:25:53 <advil> I'll close it tomorrow if there aren't any further problems that show up once it spreads to the servers
23:35:31 <hellmonk> so are there any plans to alter zin's mutation related stuff now that curemut doesnt exist
23:36:03 <hellmonk> see also https://crawl.develz.org/tavern/viewtopic.php?f=8&t=23080
23:43:43 <advil> no idea
23:43:47 -!- kogasa is now known as Doesnty
23:43:51 <minmay> hellmonk: no
23:44:12 <hellmonk> but y tho
23:45:37 <advil> maybe Brannock has thought about it?
23:46:00 <minmay> hellmonk: b/c nobody has bothered to plan that yet
23:46:34 <hellmonk> lol
23:46:36 <advil> yeah, I think it's just that all individuals who could have plans currently lack them, or at least haven't announced them
23:48:19 <advil> "it's smart to convert to Zin when you have a mutation set you want to preserve." I'm having trouble coming up with a mutation set that would be this good, I guess unless you are already with a good god
23:48:25 <advil> maybe it's more of a flavor argument
23:48:38 <minmay> who is htat a quote from
23:48:48 <advil> first post in the thread hellmonk linked
23:48:51 <minmay> because theyre silly, you can preserve a mutation set without zin 100% fine
23:49:14 <advil> not swayed by the second argument either
23:49:35 <Cheibriados> 03Doesnt02 07https://github.com/crawl/crawl/pull/485 * 0.20-a0-1032-g076eab8: fix one (1) conflict 10(4 minutes ago, 0 files, 0+ 0-) 13https://github.com/crawl/crawl/commit/076eab8c7598
23:49:48 <hellmonk> I don't think the change is actually a huge deal or whatever but zin's interaction with new !mut feels weird to me
23:49:51 <advil> the proposed changes don't sound crazy though
23:50:06 <advil> what is the interaction?  I feel like I'm missing something
23:50:14 <hellmonk> you can't quaff it
23:50:22 <hellmonk> without getting penance (I think?)
23:50:33 <hellmonk> so you can't cure muts with zin without blowing your capstone one-time removal
23:51:11 <advil> ah
23:51:32 <alexjurkiewicz> so basically when you join zin you'd want to use your capstone at full piety and then never interact wiht the mutation system again
23:51:57 <alexjurkiewicz> I feel the arguments made about rmut (it lets you opt out of what is meant to be an interesting system) apply to zin somewhat now too
23:52:30 <alexjurkiewicz> I like the idea of zin not providing a capstone or rmut, and simply removing muts ("purifying the player") over time
23:54:17 <advil> the mut removal on gift timeouts as a replacement for the capstone seems interesting (
23:55:41 <Doesnty> that seems the most reasonable to me but it does have the odd consequence of
23:55:54 <Doesnty> encouraging heavily mutated players to game their piety if it's not an always-on thing
23:55:58 <Doesnty> if they like their mut set
23:56:16 <advil> from what I understood of the rMut removal discussion the point was made that if players wanted to opt out of mutations, they should play zin
23:56:23 <alexjurkiewicz> i think gift timeout prevents that, doesn't it?
23:56:49 <hellmonk> you could lock yourself below gifting piety levels
23:56:49 <alexjurkiewicz> eg if the gift timeout starts from 0*, and every time you get piety the timer decrements, you can only keep your muts by not gaining zin piety at all
23:57:02 <hellmonk> unless you make gifting start at 0*
23:57:03 <hellmonk> but then you lose acess to sanctuary
23:57:10 <Doesnty> yeah it'd have to start from 0* i think
23:57:23 <hellmonk> imo, if players want to suboptimally not go for sanctuary in order to keep a mut set just let them
23:57:35 <hellmonk> I cannot think of a single mut combination that is better than having sanctuary
23:57:40 <Doesnty> i'd rather the temptation didn't exist
23:58:06 <Doesnty> technically you're not using sanctuary that often are you...
23:58:20 <alexjurkiewicz> yeah but when you do it's literally lifesaving