Pattern viewer for forum threads

[muzik]

KILLGLIDERS appears to just give up around 2000 generations in here:

Code: Select all

#N myrtlelife-rake-gun.rle
#O silversmith, November 7 2020
#C
#C     Yet another bootstrapping rake gun, this time with
#C many glider-reflecting still lives in a MirrorLife relative.
#C
#C Reference: https://www.conwaylife.com/forums/viewtopic.php?f=11&t=4806&p=109627#p109627
x = 246, y = 476, rule = B3-kq4j/S2-n36c
165b2o$165b2o2$139b2o4b2o$139b2o4b2o2$165b2o$165b2o3$152b3o$142b2o8bo
$142b2o9bo3$126b2o$126b2o15b2o$123b2o18b2o$124b2o$123bo4$142bobo$142b
2o$143bo7$160bo$159b2o$159bobo7$116b3o22bo$118bo21b2o$117bo22bobo2$
102bo$101bo$101b3o4$84b2o$84b2o77b2o4b2o$163b2o4b2o7$87b2o21b2o54b2o$
87b2o21b2o54b2o2$206b2o$81b2o122b2o$81b2o124bo$117b2o$117bobo$117bo6b2o
55bo$124b2o56bo34b2o$180b3o34b2o2$176b2o$176b2o$91b3o$58b2o16b2o13bo$
58b2o16b2o14bo$188b2o$85b2o101b2o$85b2o$173b2o4b2o$173b2o4b2o4$54b2o
73b3o85b2o$54b2o75bo80b3o2b2o$130bo81bo$209b2o2bo$87b2o120b2o$87b2o30b
2o$80b2o37b2o96b2o$75b3o2b2o135b2o$75bo$76bo5$72bo6$100b2o$100b2o3$
131b2o$113bobo15b3o37b2o$114b2o15b3o37b2o$114bo16b3o$130bob2o45b2o$89bo
40b3o46b2o$90b2o39bo11bo31b3o$89b2o80b2o2bo$171b2o3bo2$73b3o103b2o$
75bo24b2o77b2o$74bo25b2o$65b2o104b2o$65b2o104b2o$74bo$73bobo$57b2o14bob
o$57b2o15bo12b2o56b2o$65b2o15b2o3b2o55b2o$65b2o15b2o34b2o26bo$74bo43b
2o$35b2o9b2o25bobo13bo21bo$35b2o9b2o25bobo33b2o$74bo2b2o8b2o21b2o57bo
$65b2o11bo3b2o3b2o68b2o9bobo$65b2o9bobo3b2o29b3o41b2o9bobo$74bo38bo
52bo2bo2bo$52b2o19bobo20b2o16bo34b2o15b2o3b2o$52bobo18bobo20b2o51b2o14b
obo4bo$52bo21bo12b2o53b2o$87b2o53b2o13b2o$157b2o2$48bo$47b2o$47bobo
21b2o14b2o$71b2o14b2o2$57b2o$57b2o$99b2o$71b2o25bobo23b2o$71b2o14b2o11b
o23b2o$48bo38b2o$47bo$137b2o$55b2o28b2o50b2o$55b2o28b2o11b2o66b2o$98b
2o66b2o$41b2o5bo$41b2o4bo152b2o$200b2o13b2o$55b2o158b2o$55b2o128b2o$
185b2o25b2o$48bo32bobo127bobo$47bo33b2o130bo$82bo3$173b2o$131b2o40b2o$
131bobo31bobo$92b2o37bo33b2o$92b2o72bo14b2o$181b2o13b2o$167b2o27b2o$
40b2o22b2o93b2o6b2o$40b2o22b2o93bobo$68b2o89bo$68b2o3$49bo$49b2o29b2o$
42b2o4bobo30b2o$42b2o36bo$179bo$88b2o88b2o14b3o$52b2o34b2o88bobo13bo$
51b2o142bo$53bo$60b2o36b2o$60b2o36b2o$42b2o3bo$42b2o3b2o$46bobo$62b2o
21b2o14bo$60bo2bo21b2o37b2o$60b2o61bobo$70b2o53bo$70b2o2$62b2o$62b2o
70b2o$42b2o89b2o$42b2o70b2o19bo$60b2o51b3o$60b2o20b3o28b3o$84bo28b2obo
104b2o$83bo30b3o104b2o13b2o$42b2o71bo120b2o$42b2o164b2o3b2o$208b2o3b
2o2$131bobo$101b2o27bo61b2o36b2o$101b2o9bo17bo3bo17b2o38b2o37b2o$60b2o
51b2o15bo3bo17b2o76bo$60b2o50b2o16bo12b2o$42b2o21b2o50bobo10bo2bo9b2o
$42b2o21b2o50bob2o9b3o$119b2o$73b2o45bo$60b2o11b2o110bo31b2o$60b2o59b
3o2b2o56b2o26b2o3b2o12b2o$65b2o48bo5b3ob3o56bobo25b2o17b2o$65b2o48bo
5b3ob2obo70bo$114bo6b3o2b3o18b2o51bo$73b2o45bob2o3bo19b2o49b3o$73b2o
45b3o84b3o$42b2o77bo87bo$42b2o21b2o141bo$60b2o3b2o$50b3o7b2o$52bo20b2o
19b2o$51bo21b2o18b2o120b2o$42b2o51bo119b2o$42b2o174bo$50b2o165b2o$50bo
bo121b2o4b2o36bo$50bo123b2o5b2o8b2o$180bo10b2o$208b2o31b2o$108b2o28b
2o68b2o31b2o$48b3o57b2o28b2o$50bo181bo$49bo181bobo$231bobo$171b2o4b2o
15b2o23b2o11bo$66b2o70b2o31b2o4b2o15b2o23b2o$66b2o70b2o65b2o4b2o$105b
2o98b2o4b2o29b2o$105b2o135b2o$70b2o$70b2o$52bo13b2o5b3o88bo$52b2o12b2o
7bo87b2o51b2o$51bobo20bo88bobo50b2o13b2o$231b2o2$101b2o43b2o$101b2o
42bobo54b2o$147bo54b2o$108b2o28b2o$108b2o28b2o15b2o82bo$78b2o75b2o22bo
58b2o4b2o$79b2o34bo31b2o29b2o51b2o5bobo3b2o$78bo36bobo29b2o29bobo50b2o
$108b2o5b2o$108b2o64b2o$105b2o35bo14bo16bobo$105b2o34bobo14bo15bo24b
2o4b2o$141bobo12b3o40b2o4b2o25b2o$142bo28b2o59b2o$63bo107b2o$63b2o96bo
b2o$62bobo96b2obo51b2o$10b2o71b2o131b2o$10b2o72b2o157b2o$61b2o13b2o5b
o6b2o110b2o4b2o33b2o$18b2o2b2o38b2o12b2o12b2o110b2o4b2o$18b2o2b2o37bo
120b2o$107b2o49b2o21bobo32b2o$107b2o29b2o18b2o23bo32b2o$37bo48b2o4b2o
4b2o38b2o91bo$37b2o47b2o4b2o4b2o130bobo$36bobo74bo21bobo92bobo$106b2o
4b2o21bo4bo70b2o14b2o2bo$106bobo3bobo20b2o44b2o4b2o21bobo14bo$107bobo
28b2o41b2o4b2o23bo14bobo$108b2o28b2o$24b3o143b2o70b2o$26bo56b2o4b2o4b
2o73b2o70b2o$2o23bo57b2o4b2o4b2o$2o$3bo11b2o$2b3o10b2o48b2o4b2o4b2o4b
2o4b2o4b2o4b2o68bo$2b2o61b2o4b2o4b2o4b2o4b2o4b2o4b2o66bobo$170b2o11b
2o4b2o$183b2o4b2o$32b2o$32b2o42b3o$14b2o60bo$14b2o62bo$68b2o4b2o4b2o4b
2o4b2o4b2o21bo$68b2o4b2o4b2o4b2o4b2o4b2o19bobo$118bo67b2o$60b2o57bobo
64b2o15b2o$20b2o28b2o8b2o59bo81b2o$22bo26b2o144b2o19b3o$20bo30bo143b
2o14b2o3bo$2o111bobo42bo52b2o4bo11b2o$2o63b2o4b2o43bo41b2o69b2o$65b2o
4b2o39bo3bo10b3o27bobo$116bo12bo89b2o$79b2o32bo2bo11bo91b2o$79b2o33b3o
5b2o95bo$18b2o102bobo$18b2o5b2o96b2o97b2o$24bobo96b2ob2o33bo61b2o$26b
o18b2o21b2o54bo36b2o30b2o27bo$45b2o21b2o50b2o4bo3b2o28bobo30b2o$120bo
8b3o103b2o$108b2o11b2o6b3o6b2o95b2o$58bo49b2o19b3o6b3o$58bobo68b2obo
17b2o$58b2o62b3o5b3o17b2o50b2o$122b2o7bo70b2o$54b2o67b2o13b2o$54b2o68b
o13b2o$24b2o96bo2bobo$24b2o7bo87bobo4bo$6b2o24b2o20b2o4b2o58bo7bo$6b
2o24bobo19b2o4b2o58bo4bo2bo$120bo5b3o$25b2o93bo2bo$25b2o93b3o$49b2o49b
2o$48b2o50b2o81b2o$50bo32bo98bobo$57b2o22b2o25b2o74bo17bo$57b2o23b2o
24b2o$31b3o$6b2o23bo76b2o$6b2o24bo75b2o$9bo88b2o$8b2o88b2o62b2o$9bo
152b2o$43b2o61b2o61bobo$43b2o27b2o32bo2bo60b2o$72b2o34b2o60bo$98b2o$98b
2o2bobo$21b2o79b2o$21b2o10b2o68bo2b2o$33b2o71b2o4$95b2o$66b2o27b2o$66b
2o128b2o$184b2o9bobo45b2o$6b2o100b2o28b2o44b2o11bo45b2o$6b2o10b2o88b
2o28b2o$18b2o111bo$132bo106b3o$130b3o108bo$138b2o100bo$23b2o113b2o$23b
obo144b2o$23bo146b2o2$40b2o148b2o$40b2o112b2o34b2o2$101bo$99b2o$100b
2o$97b2o$97b2o$106b2o38b2o5b2o$106b2o38b2o5b2o$109b2o27b2o$109bobo26b
2o21b2o16b2o$33b2o6b2o66bo51b2o16b2o$33b2o6b2o$153b2o$138b2o13b2o$138b
2o$161b2o$161b2o$40bo118bo$38b2o113b2o4b2o$39b2o112b2o4bo$85b2o$84b2o
75b2o$34b2o50bo74b2o$34b2o$153b2o$153b2o3$108b2o28b2o$108b2o28b2o$
119bo$101b2o15b2o$75b2o24b2o16bo$74b2o$76bo54bo$83b2o46b2o6b2o$83b2o45bo
bo6b2o2$119b2o51b2o$119bobo50b2o$119bo2$96b2o$96bobo16b2o$97bobo15b3o
68b2o$98b2o15b3o56bo12b2o$114bob2o2b2o64bo$106b2o6b3o3bobo$106bo2bo5b
o5bo16b2o$108b2o28b2o3$129bobo$132bo5b2o$108b2o18bo3bo5b2o$92b2o14b2o
12b2o4bo3bo$92b2o38bo$122b3o4bo2bo17bo$130b3o$92bo$91b2o29b3o$91bobo
28b3o$120b2o2bo2b2o9b2o3b2o$119b3o5b3o8b2o3b2o$119b3o4bob2o$119b3o4b3o
$119b2obo4bo$120b3o20b2o$108b2o11bo20bobo$108b2o34bo2$97b2o$97b2o$
114bo$108b2o$108b2o3$138b2o$138b2o3$138b2o3b2o$96bo41b2o3b2o2$94bo3$108b
2o$108b2o4$108b2o$108b2o23bobo$134b2o$134bo2$138b2o$138b2o!
[[ KILLGLIDERS AUTOFIT STEP 64 ]]

[rowett]

LifeViewer is now open source under the GPL-3.0 license. The repo is here.

[wirehead]

LifeViewer is now open source.

THANK YOU for releasing your amazing work to the community!!

Now that LifeViewer is on Github, if I find bugs in LifeViewer would the Github bug tracker be a more appropriate place to put them, or would you still prefer using the forums?

[rowett]

Now that LifeViewer is on Github, if I find bugs in LifeViewer would the Github bug tracker be a more appropriate place to put them, or would you still prefer using the forums?

For now please report issues here.

[muzik]

KILLGLIDERS appears to just give up around 2000 generations in here:

I've managed to narrow down the issue to the following:

Code: Select all

x = 83, y = 83, rule = B3/S23
80b3o$80bo$81bo78$b2o$2o$2bo!
[[ KILLGLIDERS ]]

Of note is the fact that the bounding box of this arrangement is even in half of all generations. There still, however, appears to be some freedom of movement. This is as far to the left as the top glider can be nudged - moving it any further to the left causes KILLGLIDERS to delete everything:

Code: Select all

x = 75, y = 83, rule = B3/S23
72b3o$72bo$73bo78$b2o$2o$2bo!
[[ KILLGLIDERS ]]

And this is as far to the right as it can go:

Code: Select all

x = 92, y = 83, rule = B3/S23
89b3o$89bo$90bo78$b2o$2o$2bo!
[[ KILLGLIDERS ]]

Are edge cases like this worth fixing (similar to this one) or are they too few and far between such that they can be safely ignored?

----

Something else I noticed recently: instead of giving the "only one alternate allowed" error, this now assumes we're talking about RuleLoader:

Code: Select all

x = 16, y = 10, rule = B123/S12345678|B123/S12345678|B123/S12345678|B123/S12345678|B123/S12345678|B123/S12345678|B2c/S
o13bo13bo13bo13bo13bo13bo13bo13bo!

[muzik]

Possible further room for improvement here: this takes 224.6 seconds to recognise periodicity in Identify, and 389.0 to calculate everything. Frames get pretty bad, even at 1x, before we're even halfway through.

Code: Select all

x = 24, y = 25, rule = R6,C5,M1,S59..62,B12..17,NN
5$16.D$8.D$9.D6.2D$6.D12.D$5.C3DB6.B3D$5.2C12DC$5.2CB8D2B2C$6.2C4A7CB
$5.2B11C2B$5.6B2C4.2BA$5.2AB.9B2A$6.2A2.2A4B4A$7.12A$10.3A!
[[ SHOWTIMING EXTENDEDTIMING ]]

[rowett]

Possible further room for improvement here: this takes 224.6 seconds to recognise periodicity in Identify, and 389.0 to calculate everything. Frames get pretty bad, even at 1x, before we're even halfway through.

Code: Select all

x = 24, y = 25, rule = R6,C5,M1,S59..62,B12..17,NN
5$16.D$8.D$9.D6.2D$6.D12.D$5.C3DB6.B3D$5.2C12DC$5.2CB8D2B2C$6.2C4A7CB
$5.2B11C2B$5.6B2C4.2BA$5.2AB.9B2A$6.2A2.2A4B4A$7.12A$10.3A!
[[ SHOWTIMING EXTENDEDTIMING HISTORYSTATES 0 ]]

Remove the history states and it completes in 10 seconds on my desktop machine.

[rowett]

I've managed to narrow down the issue to the following:

Build 1116 fixes these specific cases. However more work is needed to fix the original test case pattern.

KILLGLIDERS marks gliders on the pattern boundary that are heading away as potential kills. In then looks to see if there are alive cells on the glider diagonal since these cells may be a spaceship that it will hit later. If the detected cells are another glider with the same orientation then it can kill the current one (since they are moving in parallel and will never hit).
The issue is that the detected cells are only matched as a glider if it is on the pattern boundary. This needs to be improved so it will match with gliders anywhere in the pattern. This fix is on the backlog.

Something else I noticed recently: instead of giving the "only one alternate allowed" error, this now assumes we're talking about RuleLoader:

Code: Select all

x = 16, y = 10, rule = B123/S12345678|B123/S12345678|B123/S12345678|B123/S12345678|B123/S12345678|B123/S12345678|B2c/S
o13bo13bo13bo13bo13bo13bo13bo13bo!

Fixed in build 1116. Thanks!

[muzik]

I've managed to narrow down the issue to the following:

Build 1116 fixes these specific cases.

Is build 1116 currently unreleased? Neither conwaylife.com nor catagolue appear to host it, and it isn't listed in the changelog.

There's another oddity with KILLGLIDERS I noticed which may or may not be covered by that fix: the following 90-degree rotating glider gun would ideally be recognised with a mod a quarter of its period under KILLGLIDERS, but it's instead recognised as half (180 degree rotation). The issue appears to be directional, as rotating the gun 90 degrees and identifying it again gets us the same period map as before without a 90-degree rotation.

Code: Select all

x = 6, y = 6, rule = B2n3-ky4j78/S23-cq5j
3bo$2b3o$bo3bo$2obobo$b4o$2bo!

Code: Select all

x = 6, y = 6, rule = B2n3-ky4j78/S23-cq5j
3bo$2b3o$bo3bo$2obobo$b4o$2bo!
[[ KILLGLIDERS ]]

And another thing I want to point out: saving patterns gets rid of extra spaces after #C. As Golly's pattern collection uses extra spaces like this I'd assume that their presence is canonical and LifeViewer should not attempt to collapse this.

Code: Select all

#N 4c-273954-slow-ship.rle
#O Connor Steppie, 6 February 2019
#C     The following is a spaceship which moves orthogonally with a
#C displacement of four cells every 273 954 generations, or
#C approximately one cell every 68488.5 generations.
#C This makes this spaceship the slowest known spaceship as of 2021,
#C excluding spaceships made with universal construction technology,
#C rules with adjustable-speed spaceship families, or custom rules
#C specifically created to harbor slow spaceships.
#C
#C Reference:
#C   https://www.conwaylife.com/forums/viewtopic.php?p=69941#p69941
x = 100, y = 100, rule = R45,C2,S2884-4922,B2500-4973
40b20o$36b9o10b9o$34b7o5b8o5b7o$32b6o6b12o6b6o$30b6o6b16o6b6o$28b6o5b
22o5b6o$27b5o5b26o5b5o$25b6o3b32o3b6o$24b5o4b34o4b5o$22b6o3b38o3b6o$
21b6o2b42o2b6o$20b6o2b44o2b6o$19b6o2b46o2b6o$18b6ob50ob6o$16b7ob52ob7o
$15b7ob54ob7o$15b6ob56ob6o$14b6ob58ob6o$13b6ob60ob6o$12b6ob62ob6o$11b
6ob64ob6o$10b6ob66ob6o$9b6ob68ob6o$9b82o$8b84o$7b5ob74ob5o$7b4ob76ob4o
$6b4ob78ob4o$5b90o$5b43o4b43o$4b4ob34o14b34ob4o$4b37o18b37o$3b37o20b
37o$3b3ob32o22b32ob3o$3b34o26b34o$2b4ob29o28b29ob4o$2b33o30b33o$2b32o
32b32o$b33o32b33o$b32o34b32o$b31o36b31o$b31o36b31o$31o38b31o$31o38b31o
$31o38b31o$30o40b30o$30o40b30o$30o40b30o$30o40b30o$31o38b31o$31o38b31o
$31o38b31o$31o38b31o$3ob27o38b27ob3o$3ob28o36b28ob3o$b3ob27o36b27ob3o$
b3ob27o36b27ob3o$b2o3b27o34b27o3b2o$b3o2b28o32b28o2b3o$b3o3b27o32b27o
3b3o$2b2o3b28o30b28o3b2o$2b3o2b29o28b29o2b3o$2b3o3b29o26b29o3b3o$3b3o
2b30o24b30o2b3o$3b3o3b30o22b30o3b3o$4b3o2b32o18b32o2b3o$4b3o3b32o16b
32o3b3o$5b3o2b34o12b34o2b3o$5b3o3b78o3b3o$6b3o3b76o3b3o$6b4o2b76o2b4o$
7b4o2b74o2b4o$7b4o3b72o3b4o$8b4o2b72o2b4o$9b4o2b70o2b4o$9b5o2b68o2b5o$
10b5o2b66o2b5o$11b6ob64ob6o$12b6ob62ob6o$13b6ob60ob6o$13b7ob58ob7o$14b
7ob56ob7o$15b7ob54ob7o$16b7ob52ob7o$17b7ob50ob7o$18b7ob48ob7o$19b7ob
46ob7o$20b8ob42ob8o$22b7ob40ob7o$23b8ob36ob8o$24b8o2b32o2b8o$26b8o2b
28o2b8o$27b9o2b24o2b9o$29b10o2b18o2b10o$31b10o2b14o2b10o$33b34o$35b30o
$38b24o$41b18o$48b4o!

[unname4798]

I've managed to narrow down the issue to the following:

Build 1116 fixes these specific cases.

Is build 1116 currently unreleased? Neither conwaylife.com nor catagolue appear to host it, and it isn't listed in the changelog.

There's another oddity with KILLGLIDERS I noticed which may or may not be covered by that fix: the following 90-degree rotating glider gun would ideally be recognised with a mod a quarter of its period under KILLGLIDERS, but it's instead recognised as half (180 degree rotation). The issue appears to be directional, as rotating the gun 90 degrees and identifying it again gets us the same period map as before without a 90-degree rotation.

Code: Select all

x = 6, y = 6, rule = B2n3-ky4j78/S23-cq5j
3bo$2b3o$bo3bo$2obobo$b4o$2bo!

Code: Select all

x = 6, y = 6, rule = B2n3-ky4j78/S23-cq5j
3bo$2b3o$bo3bo$2obobo$b4o$2bo!
[[ KILLGLIDERS ]]

And another thing I want to point out: saving patterns gets rid of extra spaces after #C. As Golly's pattern collection uses extra spaces like this I'd assume that their presence is canonical and LifeViewer should not attempt to collapse this.

Code: Select all

#N 4c-273954-slow-ship.rle
#O Connor Steppie, 6 February 2019
#C     The following is a spaceship which moves orthogonally with a
#C displacement of four cells every 273 954 generations, or
#C approximately one cell every 68488.5 generations.
#C This makes this spaceship the slowest known spaceship as of 2021,
#C excluding spaceships made with universal construction technology,
#C rules with adjustable-speed spaceship families, or custom rules
#C specifically created to harbor slow spaceships.
#C
#C Reference:
#C   https://www.conwaylife.com/forums/viewtopic.php?p=69941#p69941
x = 100, y = 100, rule = R45,C2,S2884-4922,B2500-4973
40b20o$36b9o10b9o$34b7o5b8o5b7o$32b6o6b12o6b6o$30b6o6b16o6b6o$28b6o5b
22o5b6o$27b5o5b26o5b5o$25b6o3b32o3b6o$24b5o4b34o4b5o$22b6o3b38o3b6o$
21b6o2b42o2b6o$20b6o2b44o2b6o$19b6o2b46o2b6o$18b6ob50ob6o$16b7ob52ob7o
$15b7ob54ob7o$15b6ob56ob6o$14b6ob58ob6o$13b6ob60ob6o$12b6ob62ob6o$11b
6ob64ob6o$10b6ob66ob6o$9b6ob68ob6o$9b82o$8b84o$7b5ob74ob5o$7b4ob76ob4o
$6b4ob78ob4o$5b90o$5b43o4b43o$4b4ob34o14b34ob4o$4b37o18b37o$3b37o20b
37o$3b3ob32o22b32ob3o$3b34o26b34o$2b4ob29o28b29ob4o$2b33o30b33o$2b32o
32b32o$b33o32b33o$b32o34b32o$b31o36b31o$b31o36b31o$31o38b31o$31o38b31o
$31o38b31o$30o40b30o$30o40b30o$30o40b30o$30o40b30o$31o38b31o$31o38b31o
$31o38b31o$31o38b31o$3ob27o38b27ob3o$3ob28o36b28ob3o$b3ob27o36b27ob3o$
b3ob27o36b27ob3o$b2o3b27o34b27o3b2o$b3o2b28o32b28o2b3o$b3o3b27o32b27o
3b3o$2b2o3b28o30b28o3b2o$2b3o2b29o28b29o2b3o$2b3o3b29o26b29o3b3o$3b3o
2b30o24b30o2b3o$3b3o3b30o22b30o3b3o$4b3o2b32o18b32o2b3o$4b3o3b32o16b
32o3b3o$5b3o2b34o12b34o2b3o$5b3o3b78o3b3o$6b3o3b76o3b3o$6b4o2b76o2b4o$
7b4o2b74o2b4o$7b4o3b72o3b4o$8b4o2b72o2b4o$9b4o2b70o2b4o$9b5o2b68o2b5o$
10b5o2b66o2b5o$11b6ob64ob6o$12b6ob62ob6o$13b6ob60ob6o$13b7ob58ob7o$14b
7ob56ob7o$15b7ob54ob7o$16b7ob52ob7o$17b7ob50ob7o$18b7ob48ob7o$19b7ob
46ob7o$20b8ob42ob8o$22b7ob40ob7o$23b8ob36ob8o$24b8o2b32o2b8o$26b8o2b
28o2b8o$27b9o2b24o2b9o$29b10o2b18o2b10o$31b10o2b14o2b10o$33b34o$35b30o
$38b24o$41b18o$48b4o!

Build 1116 is released.

[rowett]

There's another oddity with KILLGLIDERS I noticed which may or may not be covered by that fix: the following 90-degree rotating glider gun would ideally be recognised with a mod a quarter of its period under KILLGLIDERS, but it's instead recognised as half (180 degree rotation).

[[ KILLGLIDERS ]] is not supported with Identify since it doesn't guarantee when a glider will be killed.

And another thing I want to point out: saving patterns gets rid of extra spaces after #C. As Golly's pattern collection uses extra spaces like this I'd assume that their presence is canonical and LifeViewer should not attempt to collapse this.

Fixed in build 1117. Only trailing whitespace is now removed.

[muzik]

There's another oddity with KILLGLIDERS I noticed which may or may not be covered by that fix: the following 90-degree rotating glider gun would ideally be recognised with a mod a quarter of its period under KILLGLIDERS, but it's instead recognised as half (180 degree rotation).

[[ KILLGLIDERS ]] is not supported with Identify since it doesn't guarantee when a glider will be killed.

The issue isn't with Identify it's with KILLGLIDERS, I was just using Identify to make the issue clearer.

It'd probably be more obvious if we use a different theme as to highlight the envelope, which should be 90 degree rotationally symmetric but is 180 degree instead:

Code: Select all

x = 6, y = 6, rule = B2n3-ky4j78/S23-cq5j
3bo$2b3o$bo3bo$2obobo$b4o$2bo!
[[ THEME Red STARTFROM 1600 ZOOM 4 ]]

Code: Select all

x = 6, y = 6, rule = B2n3-ky4j78/S23-cq5j
3bo$2b3o$bo3bo$2obobo$b4o$2bo!
[[ KILLGLIDERS THEME Red STARTFROM 1600 ZOOM 4 ]]

[rowett]

The issue isn't with Identify it's with KILLGLIDERS, I was just using Identify to make the issue clearer.

OK. It was covered under the earlier fix.

[muzik]

Here's a technical feature request: could something be added in Help, or somewhere else, which lists the color associated with each "internal" state?

It was previously possible to see this via a bug of sorts, in which you'd provide a pattern with more states than the rule expects. However, this method broke in build 1026 due to changes to invalid state handling.

Code: Select all

x = 16, y = 16, rule = B3/S23History
.ABCDEFGHIJKLMNO$PQRSTUVWXpApBpCpDpEpFpG$pHpIpJpKpLpMpNpOpPpQpRpSpTpU
pVpW$pXqAqBqCqDqEqFqGqHqIqJqKqLqMqNqO$qPqQqRqSqTqUqVqWqXrArBrCrDrErFrG
$rHrIrJrKrLrMrNrOrPrQrRrSrTrUrVrW$rXsAsBsCsDsEsFsGsHsIsJsKsLsMsNsO$sP
sQsRsSsTsUsVsWsXtAtBtCtDtEtFtG$tHtItJtKtLtMtNtOtPtQtRtStTtUtVtW$tXuAuB
uCuDuEuFuGuHuIuJuKuLuMuNuO$uPuQuRuSuTuUuVuWuXvAvBvCvDvEvFvG$vHvIvJvKvL
vMvNvOvPvQvRvSvTvUvVvW$vXwAwBwCwDwEwFwGwHwIwJwKwLwMwNwO$wPwQwRwSwTwUwV
wWwXxAxBxCxDxExFxG$xHxIxJxKxLxMxNxOxPxQxRxSxTxUxVxW$xXyAyByCyDyEyFyGyH
yIyJyKyLyMyNyO!

The above pattern would produce the following result:


I'd be interested in seeing what exact color values are assigned to each specific history/age state, especially since we can now see the advanced states' internal numbering via Settings > Advanced > State Numbers.

Here's the pattern in a valid rule:

Code: Select all

x = 16, y = 16, rule = //256
.ABCDEFGHIJKLMNO$PQRSTUVWXpApBpCpDpEpFpG$pHpIpJpKpLpMpNpOpPpQpRpSpTpU
pVpW$pXqAqBqCqDqEqFqGqHqIqJqKqLqMqNqO$qPqQqRqSqTqUqVqWqXrArBrCrDrErFrG
$rHrIrJrKrLrMrNrOrPrQrRrSrTrUrVrW$rXsAsBsCsDsEsFsGsHsIsJsKsLsMsNsO$sP
sQsRsSsTsUsVsWsXtAtBtCtDtEtFtG$tHtItJtKtLtMtNtOtPtQtRtStTtUtVtW$tXuAuB
uCuDuEuFuGuHuIuJuKuLuMuNuO$uPuQuRuSuTuUuVuWuXvAvBvCvDvEvFvG$vHvIvJvKvL
vMvNvOvPvQvRvSvTvUvVvW$vXwAwBwCwDwEwFwGwHwIwJwKwLwMwNwO$wPwQwRwSwTwUwV
wWwXxAxBxCxDxExFxG$xHxIxJxKxLxMxNxOxPxQxRxSxTxUxVxW$xXyAyByCyDyEyFyGyH
yIyJyKyLyMyNyO!

Another very minor issue I've noticed recently is that Select and Paste boxes appear to be offset by a pixel or so once again.

[rowett]

Here's a technical feature request: could something be added in Help, or somewhere else, which lists the color associated with each "internal" state?

The history and alive states are just linear interpolations between the relevant Theme colours. So if you know how many history or age states there are it's simple to calculate.

Another very minor issue I've noticed recently is that Select and Paste boxes appear to be offset by a pixel or so once again.

Please provide an example pattern.

[muzik]

Another very minor issue I've noticed recently is that Select and Paste boxes appear to be offset by a pixel or so once again.

Please provide an example pattern.

The 256-state example Generations pattern in the above post appears to be subject to this issue on my end (see the left and right edges):

[muzik]

This will likely happen when I get round to implementing 1D Totalistic rules.

I've been thinking about 1D rules lately, would it be possible to implement 1D totalistic rules as a custom neighbourhood in the HROT algorithm (N1)? Assuming you fill in all survival conditions this should work, although it might be a bit of a hacky solution and wouldn't support the MCell notation, 1D cell history processing nor 1D Generations-type rules.

Code: Select all

x = 1, y = 1, rule = R2,C2,S0-5,B1,3,5,N@0003e0
o!
[[ AUTOFIT ]]

I've also been editing some hexagonal and triangular patterns lately, and the icons switching to rectangles at 4.0 zoom can get annoying sometimes. Could hexagons and triangles be made to render out as far as zoom level 2.0 (although possibly with a toggle in case this causes extra lag)?

Turning on Frame Times also appears to provide extra debug information to the top left as of recently.

[muzik]

It appears that saving patterns in higher-range rules attempts to include dead cells, resulting in a different output RLE in what is otherwise the same pattern in the same rule. Note how the top correctly saves a 13x13 pattern, but the bottom saves a 15x15 pattern containing extra state-0 cells.

Code: Select all

x = 4, y = 3, rule = B3/S23
b2o$2obo$b2o!
[[ STARTFROM 16 ]]

Code: Select all

x = 4, y = 3, rule = R1,C2,S2-3,B3
b2o$2obo$b2o!
[[ STARTFROM 16 ]]

[rowett]

It appears that saving patterns in higher-range rules attempts to include dead cells

Fixed in build 1120. Thanks for reporting!

[rowett]

I've been thinking about 1D rules lately, would it be possible to implement 1D totalistic rules as a custom neighbourhood in the HROT algorithm (N1)?

When I get to this I'd rather implement it with the correct notation.

I've also been editing some hexagonal and triangular patterns lately, and the icons switching to rectangles at 4.0 zoom can get annoying sometimes. Could hexagons and triangles be made to render out as far as zoom level 2.0 (although possibly with a toggle in case this causes extra lag)?

No since I don't see a benefit since at that size the triangles or hexagons are so small it's not much difference and it would increase lag significantly for populated patterns.

Turning on Frame Times also appears to provide extra debug information to the top left as of recently.

True. It only appears on touch devices if Settings>Advanced>Timing Details is switched on. It will disappear when the pinch to zoom support is complete.

[muzik]

On this pattern, not all of the bounded grid edge cells are shown, and are only loaded in once a spaceship travels far enough out. If we reset to T=0 afterwards, those cells disappear, but going back to T=1 makes them instantaneously reappear. Are either of these two things expected behaviour?

Code: Select all

x = 5, y = 4, rule = B3/S23:T0,15
o2bo$4bo$o3bo$b4o!
[[ ZOOM -4 STEP 32 ]]

It also seems that autofit doesn't automatically change to fit patterns if the size/scale of the viewer popup changes. Disabling and re-enabling autofit appears to correct this issue. May be possible to reproduce on desktop by resizing the browser window.

Is Rand All used on isotropic non-totalistic rules giving outer-totalistic results intended, as opposed to a random isotropic non-totalistic rule?

Code: Select all

x = 5, y = 4, rule = B3/S23-ijn4a
o2bo$4bo$o3bo$b4o!

It also seems that when drawing or removing cells at T=0, the maximum population counter on the left side of the graph will be off by one, either one too high or too low.

Code: Select all

x = 4, y = 3, rule = B2e3cei4ceik5ikr6-an7c8/S1c2-ai3aj4aqw5ry6-ae
b2o$o2bo$b2o!
[[ GRAPH SHOWGENSTATS ]]

Once Identify detects an object's period, would it be possible for a message to be displayed before the viewer freezes? I know a progress bar or something is or was planned, but I'd prefer some sort of indicator like this in the meantime.

[rowett]

On this pattern, not all of the bounded grid edge cells are shown, and are only loaded in once a spaceship travels far enough out. If we reset to T=0 afterwards, those cells disappear, but going back to T=1 makes them instantaneously reappear. Are either of these two things expected behaviour?

The first one is expected. The grid size defaults to 512x256 and only grows when the pattern approaches the grid boundary.
The second one was a bug (now fixed). There was a difference of opinion on Reset between the now expanded grid size and the reset Snapshot. It only effected rendering.

It also seems that autofit doesn't automatically change to fit patterns if the size/scale of the viewer popup changes. Disabling and re-enabling autofit appears to correct this issue. May be possible to reproduce on desktop by resizing the browser window.

Fixed, thanks!

Is Rand All used on isotropic non-totalistic rules giving outer-totalistic results intended, as opposed to a random isotropic non-totalistic rule?

Rand All doesn't yet support isotropic non-totalistic rule generation (for any neighbourhood).

It also seems that when drawing or removing cells at T=0, the maximum population counter on the left side of the graph will be off by one, either one too high or too low.

Fixed, thanks!

Once Identify detects an object's period, would it be possible for a message to be displayed before the viewer freezes? I know a progress bar or something is or was planned, but I'd prefer some sort of indicator like this in the meantime.

Done, thanks.

[breaker's glider gun]

(actually about ruleloader, but I couldn't find a dedicated thread about it. if there is one, please move this to it.)

It would be nice to have a symmetry (in the symmety:____ line of a ruletable) that's like permute, but preserves wether something is orthogonal or diagonal.
(or, if there is a symmetry like that, please tell me about it! it would make debugging my wire rule SO much easier)

[confocaloid]

(actually about ruleloader, but I couldn't find a dedicated thread about it. if there is one, please move this to it.)

It would be nice to have a symmetry (in the symmety:____ line of a ruletable) that's like permute, but preserves wether something is orthogonal or diagonal.
(or, if there is a symmetry like that, please tell me about it! it would make debugging my wire rule SO much easier)

If I remember correctly (?) that idea was mentioned on the roadmap for Golly at some point, but was not implemented. Adding such a symmetry could simplify creating ruletables in the "intermediate rulespace" between permute-symmetric rules and isotropic rules. The future state of a cell depends on its current state, on the multiset (bag) of states of orthogonal neighbours, and on the multiset (bag) of states of diagonal neighbours. With three cellstates, there would be 3 x 15 x 15 = 675 distinguishable conditions (and 3^675 possible rules with this symmetry).

[wirehead]

(actually about ruleloader, but I couldn't find a dedicated thread about it. if there is one, please move this to it.)

It would be nice to have a symmetry (in the symmety:____ line of a ruletable) that's like permute, but preserves wether something is orthogonal or diagonal.
(or, if there is a symmetry like that, please tell me about it! it would make debugging my wire rule SO much easier)

If I remember correctly (?) that idea was mentioned on the roadmap for Golly at some point, but was not implemented. Adding such a symmetry could simplify creating ruletables in the "intermediate rulespace" between permute-symmetric rules and isotropic rules. The future state of a cell depends on its current state, on the multiset (bag) of states of orthogonal neighbours, and on the multiset (bag) of states of diagonal neighbours. With three cellstates, there would be 3 x 15 x 15 = 675 distinguishable conditions (and 3^675 possible rules with this symmetry).

There is the Nutshell format, which has a symmetry called nutshell.AlternatingPermute which is exactly what confocaloid has described. The nutshell format is a bit harder to work with if you are already used to editing raw rule tables by hand, though.