Pattern viewer for forum threads

[muzik]

Brilliant! Always great to see more rule families get support and be made more accessible to everyone (even if the lack of golly and lifelib support is a considerable handicap for their research, but LifeViewer's editing tools are definitely a step towards making that less of a concern).

I've been testing to see if I can get some other LifeViewer bugs (on the same hardware as described before) to show up consistently so they can be patched out of existence once and for all. I've found a way to get one minor yet annoying bug where the green selection box doesn't show up immediately when it should, but it isn't completely consistently reproducible (shows up roughly 30% of the time with the following steps):

1. Go to this post viewtopic.php?f=3&p=98835#p98806 making sure to refresh after each attempt
2. Use the |<| button to revert back to gen 0 and disable the autostart/autofit parameters
3. Hit the select menu
4. Hit Select All button - the green box should show up most of the time, but occasionally it doesn't until the viewer is updated by another action
Optional 5. Hit the button that sets all selected off cells to on and vice versa. It seems as though this far more frequently requires the viewer to be updated by a subsequent action to take effect.

Also it seems like the number of aliases in OCA has exploded over the past few months, so I'll try and see if I can consolidate those into a list that can be added to LifeViewer in the coming weeks.

[GUYTU6J]

Great job for the recent update. Can we write more commands for randomization so that every time a lifeviewer pops up it will show a different random initial configuration in a fixed rule? The required script would look like this: (the command names are just for example, it can be something else)
[cod]x = 0, y = 0, rule = 345/2/4
!
#C [[ RANDOMIZE RANDT <> RANDWIDTH <> RANDHEIGHT <> RANDX <> RANDY <> RANDFILL <> RANDSTATECHANCE <> <>]]
[/cod]
Which means to RANDOMIZE a RANDWIDTH-by-RANDHEIGHT region whose upper left corner is at (RANDX,RANDY) (default (0,0)) with RANDFILL percentage (cells with non-zero states versus all cells) at generation RANDT (default 0). RANDSTATECHANCE n p tells the percentage p of state n cells among all non-zero-state cells, which by default should be p=100 for n=1 and p=0 for else.
If completed, we can embed a viewer in LifeWiki Template:Rule but that's the topic for another thread.

[muzik]

Are there any plans to update this version of the applet? https://lazyslug.com/lifeview/

[rowett]

I've found a way to get one minor yet annoying bug where the green selection box doesn't show up immediately when it should, but it isn't completely consistently reproducible

The latest build attempts to fix this but I can't test it. Please let me know.

[rowett]

Are there any plans to update this version of the applet? https://lazyslug.com/lifeview/

No. I may replace that page with the plugin at some stage.

[rowett]

Can we write more commands for randomization so that every time a lifeviewer pops up it will show a different random initial configuration in a fixed rule?

Yes I have added a [[ RANDOMIZE ]] script command that generates a random pattern on load. Most of the other commands you mentioned already exist. For more details see Help->Scripts->Randomize.

Code: Select all

#C [[ RANDOMIZE ]]
x = 1, y = 1, rule = B3S23
!

[muzik]

I've found a way to get one minor yet annoying bug where the green selection box doesn't show up immediately when it should, but it isn't completely consistently reproducible

The latest build attempts to fix this but I can't test it. Please let me know.

I can't say for sure but I definitely seem to be getting a lot less (if any) of the annoying delayed/stuck buttons in general now in the viewer, which is another thing that's been bugging me for a while now. Fantastic!

The only remaining thing that comes to mind is that the performance throttling blue box occasionally returns to the viewer window, but I have absolutely no idea on how to actually reproduce this, it just happens from time to time.

[muzik]

Another long-standing annoyance is that sometimes, when in the browser, pressing the home button then opening safari again sometimes causes the viewer to jump all the way to the left of the screen instead of it being at the right where it should be.

I've sent in a commit to catagolue with several new aliases, so it should only be a matter of time before they get accepted in. Don't forget to add these ones in the meantime though: https://gitlab.com/apgoucher/catagolue/ ... 4c9a6d04f8

[rowett]

Another long-standing annoyance is that sometimes, when in the browser, pressing the home button then opening safari again sometimes causes the viewer to jump all the way to the left of the screen instead of it being at the right where it should be.

I'm not going to fix this since there's no way I can reproduce.

Don't forget to add these ones in the meantime though: https://gitlab.com/apgoucher/catagolue/ ... 4c9a6d04f8

Done, thanks.

[muzik]

Can LifeViewer use the @NAMES field in rule tables to get names for each state that can be used to specify custom colours for cells and also display when selected in the editing menu?

Code: Select all

x = 13, y = 13, rule = WireWorld
13C$.11C$.11C$.11C$.11C$.11C$.5CA5C$.11C$.11C$.11C$.11C$.11C$13C!
[[ COLOR "copper wire" 255 255 0 ]]

Identify on the above pattern also gives a NaN value

I'm also still encountering the bug with selections not always immediately showing options.

[muzik]

Is "New pattern" supposed to show up when state 6 cells are present in LifeHistory? They definitely have more functionality than being just dead cells so this seems wrong to me. Not sure if this would also apply to states 2 and 4.

Code: Select all

x = 22, y = 23, rule = LifeHistory
15.2F$13.2F2.F$11.2F4.F$10.F7.F$9.F9.F$3.F5.F10.F$.2F.F3.F12.F$.F2.F
3.F12.F$F3.F2.F11.3F$F4.F.F6.5F$F5.F2.5F$F5.3F$F5.F$F5.F$F5.F$F5.F$F
5.F.F$F4.3F$.4F3.F$9.F$10.2F$11.F$12.F!

[muzik]

That giant bunch of aliases got committed: https://gitlab.com/apgoucher/catagolue/ ... 6267deefb2

Also, a minor annoyance with Identify on mobile platforms is that it sometimes displays the explanation for the resulting value that appears in the same place as the Identify button, probably because it believes the mouse is still there.

[rowett]

That giant bunch of aliases got committed: https://gitlab.com/apgoucher/catagolue/ ... 6267deefb2

Added, thanks.

[muzik]

LifeViewer seems to freak out when randomising L2 rules from within the viewer. Probably doesn't like that the neighbourhood name is a number.

Code: Select all

x = 13, y = 14, rule = R6,C2,M1,S38..59,B32..45,N2
7b2o$5b6o$3b3ob3o$2b10o$b5o2b3o$b4o5b3o$5o5b3o$2ob2o5b3o$5o5b2o$b5o3b
3o$b9o$2bob5o$5b4o$6bo!

"Alt Gridlines" is also displayed as on in the menu even for non-Margolus rules.

Also will the UI themes I posted a while ago be added? I kinda miss the UI matching the colours of the pattern in question for the Blues theme, and having them match for every theme would be great.

[rowett]

LifeViewer seems to freak out when randomising L2 rules from within the viewer.

Fixed, thanks for reporting!

"Alt Gridlines" is also displayed as on in the menu even for non-Margolus rules.

Fixed, thanks again.

Also will the UI themes I posted a while ago be added? I kinda miss the UI matching the colours of the pattern in question for the Blues theme, and having them match for every theme would be great.

Possibly. I'll put it on the backlog.

[muzik]

Will margolus rules that birth configurations other than 0 and 15 on 0 cells be supported on bounded grids? MCell can support these just fine, and since LifeViewer also supports strobing B0 on bounded grids for HROT this doesnt seem impossible.

Code: Select all

x = 18, y = 18, rule = M1,0,4,3,2,5,9,7,1,6,10,13,12,11,15,14
18o$18o$18o$18o$18o$18o$18o$18o$18o$18o$18o$18o$18o$18o$18o$18o$18o$
18o!

The following demonstration also appears to have broken at some point: viewtopic.php?f=11&t=4103#p82510

And Margolus Identify needs serious help:

Code: Select all

 x = 2, y = 6, rule = M0,2,8,5,1,12,3,7,4,9,10,11,6,13,14,15
o$bo3$bo$bo!

[rowett]

Will margolus rules that birth configurations other than 0 and 15 on 0 cells be supported on bounded grids?

No, I don't like strobing.

The following demonstration also appears to have broken at some point: viewtopic.php?f=11&t=4103#p82510

Fixed, thanks!

And Margolus Identify needs serious help

True. Identify doesn't support Margolus yet.

[muzik]

Code: Select all

x = 13, y = 14, rule = R6,C2,M1,S38..59,B32..45,N2
7b2o$5b6o$3b3ob3o$2b10o$b5o2b3o$b4o5b3o$5o5b3o$2ob2o5b3o$5o5b2o$b5o3b
3o$b9o$2bob5o$5b4o$6bo!

Why I hadn't reported this a year back is beyond even me (maybe editing was too early in development?), but the Cut/Copy functionality seems completely broken on mobile platforms (with sync active), instead unleashing an Unholy Trinity of bugs upon the world:

- The page immediately scrolls up to the top, likely for the same reason as that other case I reported;
- A performance-throttling blue box appears around the viewer window, and
- Nothing is actually copied to the clipboard in the first place.

I don't know if it's actually worth fixing this at all, since this is probably more apple's fault than yours, so instead I'd recommend an alternate functionality (either by using the Copy button twice in quick succession, or via a new button) that causes a text box of sorts to open up like with changing the rule, containing the RLE of the selected region that can be copied or pasted out manually. (Maybe a similar approach could be used to allow for pasting external RLEs into the viewer?)

[bubblegum]

Life ended at:

Code: Select all

x = 6, y = 1, rule = b3s23
6o!

runs generations fine

Code: Select all

x = 6, y = 1, rule = g2r1b0t3s2t3:T6,5
6o!

stops strobing

Code: Select all

x = 6, y = 1, rule = M0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
6o!

stops advancing generations

[muzik]

Code: Select all

x = 6, y = 1, rule = M0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
6o!

stops advancing generations

They definitely still count up in the T menu though, which is odd. I assume that's the number that the scripts use, as opposed to the one used to account for reversible CA?

Speaking of the T menu, this inconsistency still exists:

Code: Select all

x = 3, y = 3, rule = PCA_4
H!
[[ REVERSESTART ]]

Code: Select all

 x = 3, y = 3, rule = MargSingRot
o!
[[ REVERSESTART ]]

[rowett]

Please note: you will need to refresh your browser to view this pattern (Ctrl-F5 on Chrome).

Code: Select all

x = 138, y = 124, rule = R50,C0,M1,S2125..3702,B2125..2872,NH:T720,720
34b16o38b16o$31b30o16b30o$28b82o$26b86o$25b88o$23b92o$22b94o$20b98o$
19b100o$18b102o$17b104o$16b106o$15b108o$14b110o$13b112o$12b114o$11b
116o$11b116o$10b118o$9b120o$8b122o$8b122o$7b124o$7b124o$6b126o$5b128o$
5b128o$4b130o$4b130o$3b132o$3b132o$3b132o$2b134o$2b134o$2b134o$b136o$b
136o$b136o$b136o$138o$138o$138o$66o6b66o$61o16b61o$58o22b58o$55o28b55o
$53o32b53o$52o34b52o$50o38b50o$49o40b49o$49o40b49o$b48o40b48o$b47o42b
47o$b47o42b47o$b46o44b46o$b46o44b46o$2b45o44b45o$2b45o44b45o$2b45o44b
45o$3b43o46b43o$3b43o46b43o$3b43o46b43o$4b42o46b42o$4b42o46b42o$5b41o
46b41o$5b41o46b41o$5b41o46b41o$6b40o46b40o$6b40o46b40o$6b40o46b40o$7b
39o46b39o$7b39o46b39o$7b39o46b39o$8b38o46b38o$8b38o46b38o$9b37o46b37o$
9b37o46b37o$9b37o46b37o$10b36o46b36o$10b36o46b36o$10b36o46b36o$11b35o
46b35o$11b35o46b35o$11b35o46b35o$12b34o46b34o$12b34o46b34o$12b34o46b
34o$12b34o46b34o$13b33o46b33o$13b33o46b33o$13b33o46b33o$13b33o46b33o$
14b32o46b32o$14b32o46b32o$14b32o46b32o$14b32o46b32o$15b31o46b31o$15b
31o46b31o$15b31o46b31o$16b30o46b30o$16b30o46b30o$16b30o46b30o$17b29o
46b29o$17b29o46b29o$17b29o46b29o$18b29o44b29o$18b33o36b33o$19b37o26b
37o$19b100o$20b98o$21b96o$21b96o$22b94o$23b92o$24b90o$26b86o$28b82o$
30b78o$33b72o$37b64o$41b56o$46b46o$52b34o$58b22o!

[muzik]

Excellent! This also completes pretty much every extensible neighbourhood I can think of that's been suggested before (granted there's plenty more that could be added, but I think what we have is enough for now).

Can a button be added to squish the vertical size of hexagonal cells down slightly to make them completely regular? Them being slightly taller than they are big in the other two directions annoys me more than it really should.


Also some new aliases got added: https://gitlab.com/apgoucher/catagolue/ ... s/46/diffs

There's three more two-state aliases from MCell which I don't know the hensel/MAP notation for:

- LogicRule (isotropic)
- Snowflakes (non-isotropic)
- Sierpinski (non-isotropic)

I'm not sure how to easily figure out the rulestrings for these though.

[muzik]

Are there any plans to bring the Advance/Advance Outside selection options from Golly?

[rowett]

Are there any plans to bring the Advance/Advance Outside selection options from Golly?

Yes it's already there but only via keyboard shortcuts. See Help->Keys->Edit.

[muzik]

Could those be added as in-viewer buttons then for parity with Golly? A 180-degree rotate button could also be helpful.

Speaking of Golly, I've noticed some weird behaviour with copying stuff from bounded grids. A completely centered pattern can be copied out of Golly as so:

https://media.discordapp.net/attachment ... image0.png

yet when such a pattern is run in LifeViewer it appears offset to the bottom right by a lot:

Code: Select all

 x = 100, y = 100, rule = R5,C0,M0,S34..58,B0..0,NM:T100,100
40$41b2ob4obo2b3ob4o$40b3o3bobo3bobo3b2o$40bo5bo2bo2b4o2b2o$44bo2b2o4b
o3b3o$41b3ob2o3b3o3bobo$41bobo2bob3o2bo2bob2o$41b4obo2bob2o6bo$40bo3b
2obo3b2o2b2ob2o$40b2obob4obo2b2obobo$42bobo2b5ob2o$40b2o3bo3bobo3bobo$
40b3obobobo2b3ob2o2bo$42bobobo2bobo3b2o$41b2o2b4obo2bo3bo$42bob4obobo
2bob4o$41bobo2bo4b3obobobo$45b3obo5bo3bo$40bobob2obobobo2bob2obo$46bob
ob2o4bo$40bobobo2b4obo2b3obo!