News
What is #NordicBots?
Documentation
The Crew
Former Crew

 
 
Sign Up
Downloads
Login
Request Password

 
 
Request a bot
Available bots
Natasha Bots
Channel Information
Channel Statistics
Channel Logs
Request Information
Channel Violations
Channel Suspends

 
 
Homepage
Service Stats
Service Graphs

 
 
Rules
Q FAQ

    
 
 
  Channel logs for #lix on Wednesday the 27th of April 2016


Logs for #lix on QuakeNet

Wednesday the 27th of April 2016





[last]
Times and dates are in Central European Summer Time (GMT + 2)


[00:30:34] *** Quits: IchoTolot ([email protected]) (Signed off)
[01:01:09] *** SimonNa is now known as SimonN
[02:54:39] <SimonN> gnar schnurr schnurr, assertion failure
[02:55:08] <SimonN> from the exciting new file list with scrollbar. The list needs hotkeys to move up and down, or geoo will rage.
[02:55:15] <Proxima> :P
[02:57:45] <SimonN> I write good code, I write good code, I write good code, the crash is to remind me how I write good code that self-checks
[02:58:55] *** Quits: mobius1 ([email protected]) (Ping timeout)
[03:14:19] *** Quits: Wuzzy ([email protected]t.de) (Ping timeout)
[03:23:52] *** SimonNa ([email protected]) has joined #lix
[03:23:53] *** Q sets mode: +o SimonNa
[03:24:05] *** mobius1 ([email protected]) has joined #lix
[03:25:33] *** Wuzzy ([email protected]) has joined #lix
[03:26:27] *** Quits: Proxima ([email protected]) (Signed off)
[03:29:07] *** Quits: SimonN ([email protected]) (Ping timeout)
[04:21:36] *** SimonNa is now known as SimonN
[04:42:28] *** Quits: mobius1 ([email protected]) (Ping timeout)
[06:19:40] *** Quits: Wuzzy ([email protected]) (Signed off)
[06:37:21] *** geoo ([email protected]) has joined #lix
[06:37:21] *** Q sets mode: +o geoo
[06:57:13] <SimonN> geoo: hotkey up/down in Lix level browser works for files and dirs now
[06:57:44] <geoo> You have both dirs and files in the same list now, right?
[06:57:49] <SimonN> yes
[06:57:55] <geoo> sounds great :D
[06:58:13] <SimonN> yes yes, some bugs still, and the scrollbar is inflexible
[06:58:21] <geoo> screw the scrollbar :P
[06:58:22] <SimonN> but keyboard nav looks perfect already
[06:58:42] <SimonN> the scrollbar is good for you, so you can estimate how many levels are in the dir from size of the thumb
[06:59:08] <geoo> that's a fair point. And it's useful for Mausschubser
[06:59:13] <SimonN> yeah
[06:59:36] <SimonN> right now, they have to aim for the little buttons at the scrollbar's end. I want a click right in the track to work, too
[07:00:25] <SimonN> that's what we get for shunning GUI libs and rolling our own widgets
[07:00:31] <geoo> do you want clicking on the track to go to the point in the list corresponding the the position of the click within the srcollbar, or go down one page?
[07:01:08] <SimonN> I'm not sure. I think clicking the track at 60 % down should scroll exactly to 60 %, but every scrollbar in the wild works by pulling the thumb towards the point
[07:01:23] <geoo> Are the dirs button twice as big as levels for visual distinction as someone proposed?
[07:01:29] <SimonN> yes
[07:01:57] <geoo> I think being able to scroll down exactly one page is useful
[07:02:07] <geoo> but goign to a specific point is useful too if you have a very long list
[07:02:17] <geoo> but for that, you could just drag the scroll bar
[07:02:57] <geoo> only issue with dragging is that the mouse sensitivity in Lix doesn't seem perfect for some reason (in A4 anyway, don't remember if the issue persists in A5)
[07:03:23] <SimonN> the specific point has a feeble advantage in how I have to implement only one method; otherwise I have to implement dragging the thumb and clicking the track
[07:03:44] <geoo> dagging the thumb is what people will do anyway
[07:03:46] <SimonN> tested A5 Lix at Icho's, the mouse sensitivity issues seem gone entirely
[07:03:57] <geoo> so if you don't implement that it will defy people's expectations
[07:04:06] <geoo> oh, that's great to hear :D
[07:04:29] <SimonN> yes yes, I was anxious to see that
[07:04:55] <SimonN> would have been a serious reason to reconsider SDL2 and do a partial rewrite
[07:05:57] <SimonN> yeah, people's expectations with the scrollbar are the best criterion
[07:17:11] <Not-b224> [LixD] SimonN pushed 4 commits to master [+3/-3/±30] https://git.io/vwVu3
[07:17:12] <Not-b224> [LixD] SimonN 613dd2a - terrain browser remembers last dir
[07:17:14] <Not-b224> [LixD] SimonN a6bd85c - Pos -> Occurrence
[07:17:15] <Not-b224> [LixD] SimonN 06f7218 - main browser: dirs/files up/down by hotkey This may be bad design. gui.picker.Picker could do it. But Picker doesn't know the recent file, or the currently highlit file or directory. Idea to improve: Picker could countUntil it finds a highlit entry, then assume that's the only highlit entry, and go up/down from there. Picker should not report executeDir. Should it report
[07:17:16] <Not-b224> executeFile on mere up/down action? Would be good for the BrowserBase browsers, but not for the level editor file picker.
[07:17:17] <Not-b224> [LixD] SimonN 750e2ee - 0.2.49, terrain browser fills screen space better
[07:29:31] <SimonN> quick opinion: should a platformer safely travel beneath a buzzsaw when a walker gets killed? I do the check depending on the eye in the spritesheet, and the platformer crouches
[07:37:14] <geoo> it'd make sense as the platformer is crouched, it probably doesn't make much of a difference to gameplay unless in the rare instances when the player builds a path only to realize it's not viable for the crows
[07:37:58] <SimonN> hm
[07:38:37] <SimonN> if we keep this, should I read from the spritesheet on app start, or hardcode the values? I don't believe I can read from the spritesheet at compile time
[07:40:11] <geoo> the eye check is a bit hackish anyway, I mean usually the spritesheet shouldn't affect physics
[07:41:14] <geoo> you could write a script that dumps the eye coords into a text file and use that at compile time...don't kno
[07:41:47] <SimonN> yeah, that would be honte
[07:42:44] <SimonN> the alternative is to check at ~ 12 lo-res above the foot. That's slightly over the platformer's eye, and below the walker's
[07:43:14] <SimonN> also, the eye is not centered, climbers have it elsewhere
[07:43:25] <SimonN> centered == x-offset from the foot is 0
[07:43:26] <geoo> I'm not sure what's best here, really. Maybe make a thread
[07:43:49] <SimonN> hmm
[07:43:53] <geoo> well it makes sure climbers can't climb up a wall with buzzsaws closeby
[07:44:46] <geoo> The eye check makes kinda sense, but you get these odd cases where you can have a laser beam through the lix' body without killing them. Striclty speaking you should check the linear path between foot and eye
[07:45:18] <geoo> Then again, lix can fit through tiny terrain gaps, so why shouldn't they fit under traps? Why have the check at all
[07:50:20] <SimonN> NL has the check only at foot, and the trigger areas (TAs) are larger to make up for that
[07:53:56] <geoo> I'm heading out to uni now
[07:54:25] *** Quits: geoo ([email protected]) (Signed off)
[08:54:08] <Not-b224> [Lix] dreamsxin forked the repository - https://git.io/vCijR
[09:22:10] <Not-b224> [LixD] SimonN edited issue #63: Scrollbar: track-clicking, car-dragging, menu hotkeys - https://git.io/vVZxA
[09:23:10] <Not-b224> [LixD] SimonN commented on issue #63: Scrollbar: track-clicking, car-dragging, menu hotkeys - https://git.io/vwVoS
[10:17:13] *** Ramond__ ([email protected]) has joined #lix
[10:40:28] <SimonN> Ramond__: about time limits, I believe you would suggest a notification at level start rather than cull?
[10:43:08] <SimonN> how strongly do you feel about this? I deem time limits less of a design problem than variable spawn interval. Time limits can be seen as a UI problem, the game doesn't notify properly
[11:00:43] *** Quits: SimonN ([email protected]) (Ping timeout)
[12:56:54] *** Quits: Ramond__ ([email protected]) (Signed off)
[15:54:12] *** SimonN ([email protected]) has joined #lix
[15:54:12] *** Q sets mode: +o SimonN
[16:23:42] *** IchoTolot ([email protected]) has joined #lix
[17:29:31] *** Proxima ([email protected]) has joined #lix
[17:54:02] <SimonN> lots of fun decisions, should I be lazy or get something to eat, or order pizza so I can be lazy. Sooo much harder than class design >_>
[17:54:33] <Proxima> :)
[17:54:46] <SimonN> I'm busy IRL this week and next week, maybe in 2.5 weeks Icho and I should have a hack night
[17:54:55] <Proxima> awesome!
[17:55:13] <SimonN> yes yes
[17:56:01] <SimonN> write boring GUI windows to set level name and starting point and stuff, not fancy GUI widgets
[17:56:23] <SimonN> if it's druding but takes time nonetheless, that's good for a hack night
[17:57:01] <SimonN> IchoTolot: start writing a bachelor thesis. Then we can do hack nights for thesis work >_>
[18:02:11] <IchoTolot> I am far away from this :P
[18:02:23] <SimonN> hurry hurry
[18:02:51] <IchoTolot> I must aquire more knowlege in Bio-CS first
[18:03:01] <IchoTolot> before even thinking about this
[18:03:08] <SimonN> hnn
[18:03:28] <IchoTolot> + finding out if this is the right path anyway or another CS branch
[18:03:34] <SimonN> lemmings are green (very biological), live, and are in the computer
[18:03:57] <IchoTolot> yes i will go to the proff with that
[18:03:57] <SimonN> how should I get work done if we don't do thesis hack nights >_>
[18:04:15] <SimonN> I'm sure they will be delighted and you will get the best grade
[18:04:42] <IchoTolot> we can do another hack night at the weekend or so I have levle ideas + maybe a java programming sheet
[18:05:13] <SimonN> in 2 or 2.5 weeks would be best for me, I'm away and busy before
[18:05:17] <IchoTolot> alright
[18:06:01] <IchoTolot> now the NL wine version works better as well ^^
[18:06:07] <SimonN> ooh
[18:06:14] <SimonN> progress on the Russian front
[18:06:22] <SimonN> s/Russian/low-level/g
[18:07:08] <IchoTolot> and I need to get forward with so many things..... making levels/the dune tileset + I also need to solve NepsterLems again for the x-th time :P
[18:07:27] <IchoTolot> backroutes backroutes so many backroutes for Nepster
[18:07:55] <SimonN> because the levels are both small and have the Proxima quality
[18:08:09] <Proxima> the... what now?
[18:08:32] <SimonN> the Proxima quality needs a better name, a level has the Proxima quality when you have a general-looking skillset, can solve it in several ways, yet no way is easy
[18:08:38] <Proxima> Ah :)
[18:08:47] <IchoTolot> they are often too small ---> creating backroutes
[18:09:14] <Proxima> *is touched*
[18:09:26] <SimonN> yes yes, you mentioned this on the stream yesterday
[18:10:15] <SimonN> the 8-hatch Egyptian level is too easy with 20-of-all to qualify, but with 4 of everything maybe, the level would have it
[18:10:43] <SimonN> homework assignment: choose a good name for this quality
[18:11:52] <IchoTolot> and I have a feeling that the one-screener has a difficulty ceiling the difference between Black Hole and Neutron Star is barely noticable
[18:12:58] <IchoTolot> good levels nevertheless
[18:13:52] <IchoTolot> SimonN: As I said the level comes back in the 4th rank at the near end with only 8 lems + very few skills
[18:16:14] <SimonN> yeah, and then it with several viable routes?
[18:17:49] <IchoTolot> no just one
[18:25:14] <Proxima> I already made a suggestion in the time limits thread: resource management level
[18:34:04] <namida42> appropriate, but a tad long
[18:34:06] <namida42> "resman level" :P
[18:35:50] <IchoTolot> rl
[18:36:16] <Proxima> RML
[18:36:23] <IchoTolot> let's use the fewest letters so that nobody can recognise it! :P
[18:36:42] <IchoTolot> resource management level is fine
[18:36:53] <IchoTolot> everybody knows what it means
[18:38:51] <namida42> hm, does it need to be a general looking skillset, or simply more skills than are needed to solve the level?
[18:40:12] <namida42> eg. would "A Violation Of Lemming Rights" count? it has surplus skills but it doesn't look like there's any general pattern to it; there's no extremely easy solution (obviously it isn't too hard owing to being early in the 2nd rank)
[18:40:59] <IchoTolot> maybe sth like "smart way carving level"
[18:41:16] <Proxima> Yeah, that would count. The main feature is there isn't a single intended way to solve it, but it's still not trivial
[18:41:20] <namida42> http://neolemmix.com/levelimg/lpii/0205.png save 40 of 50; RR 40; no time limit; skillset 5 / 2 / 1 / 2 / 3 / 2 / 2 / 2
[18:41:21] <IchoTolot> you've got enough skills as long as you use them a bit smart
[18:41:36] <SimonN> 'resource management' sounds like it applies to too many levels than we want it to apply to
[18:41:52] <IchoTolot> "smart carvibg level"
[18:41:55] <IchoTolot> carving
[18:42:01] <SimonN> carving level
[18:42:40] <IchoTolot> i think there is still a difference between smart and no smart
[18:42:48] <namida42> that implies an emphasis on destructive skills
[18:42:53] <SimonN> hmm
[18:43:08] <IchoTolot> the levels with skills on mass I would call carving level
[18:43:18] <namida42> skill budgeting level?
[18:43:35] <IchoTolot> where you have to think a little I would add the smart or an "advanced"
[18:43:39] <SimonN> skill budgeting again happens on too many levels than what we have in mind
[18:44:22] <Proxima> well, I certainly won't object if you want to call it "Proxima-type" :P
[18:45:40] <SimonN> that is normal in mathematics. Some mathematicians are good at flowery, excellent names, but many are not. Those name it after the author of the first thorough treatment
[18:46:35] <SimonN> good naming is exclusive
[18:46:38] <IchoTolot> "training level" as it trains often basic tricks aswell
[18:46:52] <IchoTolot> as well argh
[18:47:17] <Proxima> :D
[18:47:18] <SimonN> Is Just Dig a training level?
[18:47:31] <SimonN> is
[18:48:51] <SimonN> good naming is exclusive: humans are eager to associate what does not belong together fully. Therefore, if the name errs slightly on exclusion, people use it for too much anyway. If the name is too inclusive, everything begins to fit that name, which is bad
[18:49:09] <IchoTolot> i would call that a tutorial level
[18:49:39] <SimonN> that wasn't the question; the question is whether 'training level' is an appropriate label
[18:49:52] <SimonN> a level can be of more than one type
[18:50:00] <IchoTolot> "sparring level"?
[18:51:07] <SimonN> lovely. Maybe a little too much emphasis on training, as opposed to an emphasis that this is a serious hard problem
[18:52:01] <IchoTolot> "warm uppers" :P
[18:52:18] <IchoTolot> as they warm you up for the harder tasks
[18:52:33] <SimonN> there is a level of Proxima quality near the end of the Lix community pack. And Icho-pack last rank, last level has this quality too
[18:52:57] <SimonN> these are not about training anymore, but they satisfy the property
[18:53:15] <IchoTolot> 1st rank last level you mean?
[18:53:28] <SimonN> last rank, last level, with the pyramid in the center
[18:53:29] <IchoTolot> as last rank last level would be: "no mercy"
[18:53:36] <IchoTolot> all out
[18:54:21] <IchoTolot> nepsters monster at the end of his pack would also fall in there
[18:54:42] *** Wuzzy ([email protected]) has joined #lix
[18:55:18] <SimonN> yes yes!
[18:55:49] <SimonN> especially Nepster's monster with save 75 %. I'm not sure if Nepester's monster requiring 100 % still qualifies, maybe that has only 1 solution
[18:56:39] <Proxima> this sounds interesting, I'm looking forward to when someone LPs Nepster's pack :)
[18:56:52] <IchoTolot> "no mercy"! :P
[18:57:10] <IchoTolot> I need to burn all the backroutes out then it will be ready
[18:57:25] <IchoTolot> even if Nepster turns mad in the process :P
[18:57:30] <Proxima> I like the idea of a final level that deliberately requires less than the number it's possible to save, but is still very hard. "You only get one bash" has that too
[18:57:49] <Proxima> but to a much lesser extent -- requires 48/50 but the intended route saves 49 anyway
[19:00:17] <IchoTolot> Nepster also said that when he gives the "OK" then videos should be uploaded as no one really wants to see backroute after backroute
[19:00:34] <IchoTolot> gives also a false impression of the pack
[19:00:55] <SimonN> okay, hmm, Icho's No Mercy doesn't have the quality, because each part has an intended solution, others become backroutes
[19:01:05] <SimonN> but Nepster's level qualifies at 75 % very much
[19:01:38] <IchoTolot> 75 and 100 % are very different levels there
[19:04:00] <namida42> hm, what about a level like To The End?
[19:04:42] <IchoTolot> pic please don't have it in my head right now
[19:04:51] *** Nepster_ ([email protected]) has joined #lix
[19:05:22] <IchoTolot> speak of the devil :P
[19:05:40] <namida42> http://neolemmix.com/levelimg/lpi/0528.png
[19:06:03] <Nepster_> The devil was lured here by Simon :)
[19:06:35] <IchoTolot> classic would simply be"builderless"
[19:07:13] <SimonN> we are interested in levels that satisfy all of these: 1. they're hard, 2. they have multiple solutions, none of them is easy 3) there need not be "the" intended solution, all the equally-hard solutions are accepted variants
[19:08:57] <SimonN> http://asdfasdf.ethz.ch/~simon/etc/levels-single-lemforum-Hopeless-labyrinthofdespair.txt.png <-- important example from the lemforum pack, 3rd- or 4th-to-last level
[19:09:22] <SimonN> and Nepster_'s 100 % 20-of-everything level has this quality too, at least when 75 % is required
[19:09:53] <IchoTolot> wait I think I solved that and it wsn't too dificult
[19:09:59] <IchoTolot> in NL though
[19:10:14] <Proxima> in NL it gives 30 of everything
[19:10:15] <Nepster_> Yes, the 75% version is not really difficult.
[19:10:16] <IchoTolot> oh batter and stuff
[19:10:27] <IchoTolot> no this is a bit different
[19:10:36] <Proxima> my Lix adaptation gives *3* of everything and requires 99%
[19:10:40] <IchoTolot> NL has a more difficult version either
[19:11:17] <Proxima> yes, I understand NL has a talisman, directly inspired by my version of the level, which is 3 of everything with only the original 8 skills, but a lower save req
[19:11:28] <SimonN> the batters and jumpers are important in the Lix version, don't know if there is a route without.
[19:11:58] <namida42> yeah, NL is (in both standard and talisman variants) save 72 of 80
[19:12:09] <Proxima> the terrain will be slightly different as well -- same layout but the pieces and gaps will be slightly different in size
[19:12:47] <Proxima> so it's hard to directly compare the two, but of course both very hard :D
[19:12:56] <namida42> also, unlike the lix version, the NL version gives you bombers (but obviously not batters or jumpers)
[19:13:34] <namida42> the non-talisman version in NL gives you 30 of everything, except only 5 floaters
[19:14:11] <Proxima> the lix version does have bombers, but you can only use 1 because of the save req :)
[19:14:36] <namida42> according to the linked image it does not
[19:14:43] <namida42> possible issue with image dumper?
[19:14:49] <SimonN> bug in the image exporter
[19:15:09] <SimonN> thanks, never noticed before, very good
[19:15:22] <namida42> iirc, the 3-of-everything challenge for it on NL requires using every skill you're allowed except for one floater (though you can still use it too, to save an extra lemming)
[19:15:27] <Proxima> yeah, the level itself has 3 imploders, in my solution I use one to help set up a landing for the right hatch
[19:15:42] <Proxima> other solutions don't, it's possible to save 100%
[19:17:46] <namida42> imploder = non-fling bomber, right?
[19:18:16] <Proxima> yes
[19:19:56] <Proxima> I'm going out now, anime evening :) thanks for interesting discussion
[19:20:46] <IchoTolot> np ;)
[19:20:55] <IchoTolot> have fun
[19:41:16] *** Flopsy86_ ([email protected]) has joined #lix
[20:32:36] *** geoo ([email protected]) has joined #lix
[20:32:36] *** Q sets mode: +o geoo
[20:42:36] <SimonN> without generating the eye matrix in noninteractive mode, we have another performance boost
[20:43:18] <geoo> you mean computing the eye positions from the spritesheet?
[20:43:23] <SimonN> yeah
[20:43:50] <SimonN> verifying a bad replay still initializes the game before erroring out. old time 0.58 seconds, new time 0.19 seconds
[20:44:31] <SimonN> this will save us millions of dollars when the Koreans start playing professionally and uploading to the django db
[20:44:45] <geoo> but that's for a single replay, right? the boost only shaves off a constant amount of time
[20:45:12] <SimonN> yes, the boost is only per app run, the app can check many replays at once
[20:46:19] <SimonN> the interactive game starts faster, too. But the spritesheet must be loaded at level start, so the first level starts slightly slower
[21:24:12] <Proxima> I'm back
[23:09:00] *** mobius1 ([email protected]) has joined #lix
[23:44:38] *** Quits: geoo ([email protected]) (Signed off)
[23:45:39] <SimonN> scanning the image tree at level start takes 60 ms
[23:46:59] <SimonN> that's acceptable. Generating lix sprites from file takes 200 to 400 ms, I'd have to profile exactly
[23:48:20] <SimonN> it's possible to distribute this workload over the first few seconds of application runtime with a self-rolled yielding coroutine
[23:55:48] <SimonN> right now, I generate the sprites immediately before first spawn, this makes a noticeable pause after opening the hatch, before the first spawn


[top]

« View logs for a different channel
Mail us brunette
brunette
brunette
brunette