Lix 0.7.22 experimental

[Simon]

Hi,

Update 2017-08-22: Lix 0.7.x is obsolete -- instead, get the most recent Lix version.







Lix 0.7.x experimental build.

Download for Windows
Branch exp in lix-unstable

Keep separate from your stable Lix 0.6.x installation! Levels, tileses, physics all differ. Let's look for physics bugs in this experimental build, make levels solvable, and re-cover levels with solving replays.

Proxima, Nepster, Rubix: See my next post for level pack maintainers -- I'll list which levels were affected by tileset changes, which levels need new proofs due to subtle physics changes, ...

Physics

• Nepster's miner mask
  
• Fix #129: Floater doesn't land inside terrain and turns (Bug found by Ramond in summer 2016)
  
• Floater doesn't go flatter than tumbler. If you assign floater to a fast tumbler, after opening the umbrella, the lix can't gain horizontal flight over the tumbler anymore. Change suggested by Proxima in summer 2016, might affect solution to Steve's Endeavor.
  
• Fix #147: Digger checks inner 14 pixels for steel. Any steel there aborts. On either side of that checked area, there is one lo-res pixel where we ignore steel. (Before, the digger stopped when there was lots of steel on the left, or lots of steel on the right, or a tiny bit amount of steel on either side. That was too complicated.)
  
• Builder creates brick further ahead, like in NL. No builder backstep. Can't seal off climber ledges behind yourself anymore by building immediately after ascending.
  
• Builders and platformes double their most recent brick when they run into a blocker. Reason: This prevents holes in bridges and allows the crowd to follow.
  
• Blocker field narrower by 1 lo-res pixel on both the left and right. Blocker field is exact same as in 0.6, this is simplest. Even though it's possible to prevent crowd from following a builder, with most blocker placements, the crowd follows.
  
• Fix #199: Don't spawn lix outside of torus bounds. Immediately wrap coordinate before the first death check.
  

Tilesets

• Added Raymanni's 3 tilesets. Thanks, Raymanni, for these excellent tilesets, released in the public domain along with Lix. I should have included them in Lix over a year ago.
  
• Remade the oriental set entirely: Removed matt/oriental, remade every tile with new looks and possibly different physics, added some extra tiles. Thanks to geoo for submitting awesome bamboo poles and pebbles! This new set is simon/oriental. All included levels point to the correct new tiles, but, due to changed tile shapes, some levels aren't solvable.
  
• Removed simon/earth, simon/brick, simon/steel. Modified Hrududu slightly, the only level that used simon/steel.
  
• Removed geoo/construction/trampoline.T.
  
• Renamed amanda/occult/occult_* => amanda/occult/*. All levels point to the renamed tiles.
  
• The steam in geoo/construction has alpha channel to look good overlapping itself, and its grey is lighter.
  

-- Simon

[Simon]

For level maintainers.

Levels with oriental tiles

Please scrutinize these levels. The oriental tiles have different shapes. Some levels aren't solvable anymore because the bamboo rods have become shorter.

lemforum/Cunning/acompletelyridiculous.txt
lemforum/Daunting/juststop.txt
lemforum/Hopeless/recyclingplant.txt
lemforum/Simple/everylix.txt
lemforum/Vicious/waitwhyisthereatree.txt

nepster/5-of-all/15_ParadiseHell.txt
nepster/Moon/16_DontFeedFrogs.txt
nepster/Planet/10_TravellingLix.txt
nepster/Planet/18_ChineseLemmings.txt

rubix/Aftermath/Parallel/8Goldmine.txt
rubix/Nightmare/bangarangpeter.txt
rubix/Nightmare/flatlined.txt
rubix/Perplexing/steptothebeat.txt
rubix/Perplexing/theclairvoyant.txt
rubix/Pleasant/illgoonaheadwait.txt
rubix/Pleasant/thegarden.txt
rubix/Sinister/7up.txt
rubix/Sinister/bridgeovertroubl.txt
rubix/Sinister/treehousetrickery.txt

Other levels with failing replays

Most of these are still solvable. A chunk of replays failed because the builder places its brick further forward, i.e., there is no backstep anymore. Please see further down this thread for a list of failing replays.

Before you start covering levels anew, review the physics changes in 0.7.0. Did I do stupid things? Especially Proxima should get in touch with me about high-level physics changes. Maybe I should chat with Nepster, namida, geoo, Ramond, or whoever else loves to ponder low-level physics.

-- Simon

[Simon]

Will probably change the blocker mechanics again. The narrower 0.7.0 blocker field is responsible for about half the failing replays. Not sure how to fix. Not sure what I want.

Details in IRC chatlog

-- Simon

[ccexplore]

The narrower 0.7.0 blocker field is responsible for about half the failing replays.

TL;DR the chatlog.  So failing as in "the intended solution no longer works (!)", or merely "the replay stops working because obviously this change slightly alters the timing of any lixes that interacted with blockers, and replays are currently strictly time-based and not position-based"?

If it's the latter, is it really any surprise?  Don't get me wrong, I understand it's kind of annoying for so many replays to get trivially invalidated, but that's a different magnitude of matter than actually affecting solvability.  Also, just so it's clear for other people, it's not half of every level, just the ones where the replays are failing.

It's not ideal perhaps, but at least we don't expect this kind of change to happen all that often, maybe never again, so maybe a one-time pain is alright?

[Simon]

Uploaded 0.7.1, same download link as in first post. Only difference to 0.7.0 is the blocker behavior, now it's mostly like in 0.6. Description in this post.

So failing as in "the intended solution no longer works (!)", or merely "the replay stops working because obviously this change slightly alters the timing of any lixes that interacted with blockers, and replays are currently strictly time-based and not position-based"?

Replay stops working because obviously the change slightly alters the timing of any lixes that interact with the blocker.

Instead of narrowing the blocker field, Nepster suggested to make a hollow field, and geoo suggested to special-case the blocker field for 0.7.1. This solves the original problem that led me to narrow the blocker field.

0.6.* blocker field:     LLLLLL-RRRRRR
0.7.0 blocker field:      LLLLL-RRRRR
0.7.1 field vs. builder: LLLLL---RRRRR
0.7.1 field vs. others:  LLLLLL-RRRRRR (as in 0.6.*)

Right, it's no surprise that the outside shape affects many replays.

The goal is to always allow the crowd to follow a blocked builder A, even if we assign the blocker to B while the builder A is inside B's future blocker field. I'd like to investigate more whether we want an intrusive change or whether special-casing is OK. Maybe the builder-brick-doubling positions its duplicated brick badly and we should look into that instead.

It's not ideal perhaps, but at least we don't expect this kind of change to happen all that often, maybe never again, so maybe a one-time pain is alright?

It's probably good style to not break replays or the blocker field. We have changed the builder already. Maybe more changes to its behavior w.r.t. blockers is enough, and walker-blocker interactions can stay like in 0.6.

-- Simon

[Simon]

Uploaded 0.7.1, download in first post. See previous post for changes.

-- Simon

[geoo]

I found one issue with the special case for the builder:

If a group of lix is tightly squeezed between two blockers, you can use a builder to get a lix out. I think this might be quite exploitable in practice to separate a second worker, but either way I think this is worse than the crowd unable to follow a turned builder.
Replay attached.

[Simon]

Thanks! Behavior in 0.6 is that the builder makes a vertical stack, everybody gets free. In 0.7.1, the builder ignores the blocker and everybody else is trapped in the inside column of the field:

0.7.1 field vs. builder: LLLLL---RRRRR
0.7.1 field vs. others:  LLLLLL-RRRRRR (as in 0.6.*)
                                ^
I too think this 0.7.1 needs a revision.

-- Simon

[Simon]

I think I want the 0.6 blocker back. That blocker has no corner cases and looks good. You can't escape a builder from between two tightly placed 0.6 blockers.

It's probably OK if crowd-follows-builder only in some cases, not all cases. With the brick-doubling 0.7 builder, you can always force crowd-follows-builder even with a 0.6 blocker: Assign blocker to a lix far ahead on the recent builder brick.

Little progress on Lix these weeks beacuse speedruns and reallife and 30 degrees C outside. Eventually, I have to come forward and declare a set of physics standard.

-- Simon

[RubiX]

Nice job on the Oriental update, its so nice now.

[Simon]

For Lix 0.7.2, I've reverted the blocker field to the 0.6 blocker field of LLLLLL-RRRRRR everywhere, against everybody, including builders.

Please look at your oriental levels. The tiles look differently in 0.7.x than in 0.6. Do your levels still look reasonable?

Let's prove all levels solvable! More details in the most recent post below.

19:09 <Nepster> SimonN: Are the physics changes (more or less) final, or do you plan further changes? I am asking, because I am wondering whether it makes sense to rerecord the solution replays.
19:09 <SimonN> I'd love to keep this stable
19:09 <Nepster> Ok, thanks.
19:10 <SimonN> I've wondered only about the blocker field in the past month, everything else seemed good. And the blocker field is now the 0.6 field

Nice job on the Oriental update, its so nice now.

Thanks!

-- Simon

[RubiX]

Overly busy with being the main installer for the company I work for.   Dont think i'll get around to checking the 39 levels for a while, but i'll keep it in mind for when ive got some spare time.

[Simon]

Lix 0.7.3 experimental.

* Same physics as 0.7.2.
* NepsterLix updated and proven 100 % solvable. Thanks!
* Replay verifier in the GUI! Merely lists unproven levels, doesn't mark browser entries.
* Reworded scroll-speed option.

Proxima, Rubix: The Lix replay browser has a new button. This verifies all replays in the current directory and all subdirectories, then lists levels that still need replays. Run it over your own replay collection, or over the most recent replay collection on github (download .zip)

Overly busy with being the main installer for the company I work for.   Dont think i'll get around to checking the 39 levels for a while, but i'll keep it in mind for when ive got some spare time.

No problem -- thanks for keeping it in mind. I can start covering your easier maps, too.

I've put the replay verifier in the GUI to help you work in little chunks, to keep track of what you've already covered.

-- Simon

[Simon]

Lix 0.7.4 experimental.

Download 0.7.4 for Windows -- keep separate from the stable 0.6 installation
Branch exp in lix-unstable

* Changed a few pixels in the bonsai tree. Lixes don't get stuck in it anymore.
* Replay browser opens at recent replay or at recently-verified dir.
* Rubix's levels: Pleasant and Sinister are 100 % solvable. Small changes.

I've covered the first 2 out of Rubix's 5 ranks, Pleasant and Sinister. I've changed the oriental levels to make them solvable, preserving the spirit as much as possible. I removed an unnecessary exploder in Center of Attention because this level is save-all.

-- Simon

[Simon]

Lix 0.7.5 experimental.

* lemforum coverage: I've covered 15 levels. We need 37 more.
* A Completely Ridiculous Level's is 48 pixels narrower.
* Down-facing tumblers drown head-first. Purely cosmetic.

Proxima: What's our plan for coverage? I don't know how to solve every level. Who shall cover what? Editing levels is OK to improve them. Everybody can pitch in with replays -- see list of not-covered levels below.

I'd like to have 100 % coverage by July 8th end of July, to declare 0.7 stable, with its tilesets and physics. I'd leave out a few not-covered levels in a pinch, but it's certainly nicer to ship all.

Are you happy with the floater-tumbler physics? Cover Endeavor to get a feeling.

-- Simon