lifeviewer bug

[rowett]

Memory leak of some sort in LTL rules, especially generations ones? They just get laggier and laggier the longer you spend in a pattern.

The performance of the LtL/HROT algo is related to the area of the bounding rectangle. So if you have a pattern like the one in your signature where you have spaceships going off to infinity then it will get slower and slower. In this case you can limit the slow down by running the pattern in a bounded grid. For example:

Code: Select all

x = 74, y = 29, rule = R1,C3,S2-3,B3,NW211111112,012:P400
$56.A$56.2BA$53.A3.2B$51.B.BA2.B$52.B4.A$4.A48.B$4.BA41.B$46.A3.B.A$52.
B$53.BA$2.A3.B$.B$2.B51.B$52.A$70.BA$68.A3B$71.BA2$60.B2.A5.A$63.B3.B
.BA$59.A4.BA2.B$69.B$65.B3$66.B$65.A!

Also, if there’s an error in a rulestring, it deletes the entire rulestring, which is REALLY annoying. Not sure if that’s a bug.

In what scenario?

[b-engine]

Use the browser's built-in webpage translator to translate the entire page with LifeViewer window open, then you'll spot that the X at top left turns into 十. Use an image or finding a similar cross from Window's Character Map should fix the problem.

Screenshot 2024-12-04 210056.png (61.88 KiB) Viewed 7162 times

[rowett]

Use the browser's built-in webpage translator to translate the entire page with LifeViewer window open, then you'll spot that the X at top left turns into 十.

Fixed, thanks!

[LuveelVoom]

Memory leak of some sort in LTL rules, especially generations ones? They just get laggier and laggier the longer you spend in a pattern.

The performance of the LtL/HROT algo is related to the area of the bounding rectangle. So if you have a pattern like the one in your signature where you have spaceships going off to infinity then it will get slower and slower. In this case you can limit the slow down by running the pattern in a bounded grid. For example:
pattern

Yes. What I am describing here persists after reducing bounding rectangle size, however.

Also, if there’s an error in a rulestring, it deletes the entire rulestring, which is REALLY annoying. Not sure if that’s a bug.

In what scenario?

Pretty much any? If the rulestring is invalid, it throws an error, sends me to the null pattern, and wipes the rulestring from memory, losing me the rulegolf. It also makes long rulestring annoying to correct.

[rowett]

Yes. What I am describing here persists after reducing bounding rectangle size, however.

Please provide an example pattern that shows the slowdown and describe the steps you take that demonstrates it. Also give details of the device being used.

Pretty much any? If the rulestring is invalid, it throws an error, sends me to the null pattern, and wipes the rulestring from memory, losing me the rulegolf. It also makes long rulestring annoying to correct.

Again I need some repeatable steps that show this.

[WhiteHawk]

I am having trouble copying patterns between different clipboards, especially when it comes to transferring patterns between LifeWiki, Catagolue, and/or the forums.

The problem is that when I try to cross-copy an object from one LV to another, I can't get it to copy over or appear on the external clipboard.

EDIT: it seems to be fixed now.

[dvgrn]

I am having trouble copying patterns between different clipboards, especially when it comes to transferring patterns between LifeWiki, Catagolue, and/or the forums.

Try writing down the very specific steps of a copy/paste sequence that is causing you trouble. If you just say you're having trouble, that's too vague to allow anyone else to duplicate the problem.

[b-engine]

Again I need some repeatable steps that show this.

Open the pattern below:

Code: Select all

x = 0, y = 0, rule = B3/S234i6c
obo$b2o$bo!

Then go to Pattern > Change Rule, add an invalid character (for example "v" which will work anywhere) into the rulestring, and the pattern will disappear, but not the rulestring.

[rowett]

Then go to Pattern > Change Rule, add an invalid character (for example "v" which will work anywhere) into the rulestring, and the pattern will disappear, but not the rulestring.

This is fixed in build 1221.

[pifricted]

When LifeViewer loads PCA rules, LifeViewer crashes.

[b-engine]

When LifeViewer loads PCA rules, LifeViewer crashes.

It doesn't:

Code: Select all

x = 25, y = 25, rule = 2PCA4,0,1,2,12,4,10,9,13,8,6,5,14,3,7,11,15|2PCA4,0,4,8,12,1,10,9,7,2,6,5,11,3,13,14,15
12.D$11.B.H$12.N2$12.D7$.A21.A$H.G.H15.B.M.B$.D21.D$4.E$3.J5$12.A2$12.
K$11.B.H$12.A! 

[pifricted]

When LifeViewer loads PCA rules, LifeViewer crashes.

It doesn't:

Code: Select all

x = 25, y = 25, rule = 2PCA4,0,1,2,12,4,10,9,13,8,6,5,14,3,7,11,15|2PCA4,0,4,8,12,1,10,9,7,2,6,5,11,3,13,14,15
12.D$11.B.H$12.N2$12.D7$.A21.A$H.G.H15.B.M.B$.D21.D$4.E$3.J5$12.A2$12.
K$11.B.H$12.A! 

On my iPad:

[b-engine]

On mobile phone:

On mobile phone:

Screenshot_20241207_173718_com.android.chrome.jpg (275.35 KiB) Viewed 6966 times

[rowett]

When LifeViewer loads PCA rules, LifeViewer crashes.

Please let me know:

1. The device you are using
2. The operating system and version
3. The browser and version

[pifricted]

Please let me know:

1. The device you are using
2. The operating system and version
3. The browser and version

1. iPad mini 2 (too old?)
2. iOS 12.5.5
3. Just Safari

[unname4798]

Please let me know:

1. The device you are using
2. The operating system and version
3. The browser and version

1. iPad mini 2 (too old?)
2. iOS 12.5.5
3. Just Safari

1. Samsung Galaxy A14
2. Android 14
3. My user-agent:

Code: Select all

Mozilla/5.0 (Linux; Android 14; SM-A145F Build/UP1A.231005.007; wv) AppleWebKit/537.36 (KHTML, like Gecko) Version/4.0 Chrome/131.0.6778.104 Mobile Safari/537.36

[b-engine]

1. Currently using Huawei P10+ (Is this old enough?)
2. Presumably Android 7.1
3. Chrome 131.0.6778.104

[rowett]

1. iPad mini 2 (too old?)
2. iOS 12.5.5
3. Just Safari

Thanks. Try build 1224 and let me know if it works for you please.

[pifricted]

Thanks. Try build 1224 and let me know if it works for you please.

It works.

[pifricted]

Move right, then you will see the bug:

Code: Select all

# [[ ANGLE 180 ZOOM 3 X -4000 ]]
x = 7802, y = 1, rule = B3/S23
7802o!

[b-engine]

It wasn't fixed because I can't reproduce the problem.

You can only reproduce it by opening LifeViewer in LifeWiki or ConwayLife forums on device with only touchscreen device. What's the browser, OS and device you're using to try to reproduce the bug?

[unname4798]

What's the browser, OS and device you're using to try to reproduce the bug?

Android 14, Chrome, Galaxy A14

[rowett]

Move right, then you will see the bug

This has been fixed and will be in the next release.

EDIT: This is now released.

[LuveelVoom]

When pressing "randomize rule" in an INT rule, a OTL rule will be rolled instead.
Steps to reproduce:
1. Create a new pattern in an INT rule (for example, B2-a3/S3k4enri7-c, just to give one with a variety of symbols)
2. Press randomize rule.
3. Observe that lifeviewer miraculously pulls out one of the 262,144 OTL rules from the space of countless undecillions of INT rules.

This glitch may occur for other rulespaces as well, I haven't yet tested it.

[confocaloid]

When pressing "randomize rule" in an INT rule, a OTL rule will be rolled instead.
Steps to reproduce:
1. Create a new pattern in an INT rule (for example, B2-a3/S3k4enri7-c, just to give one with a variety of symbols)
2. Press randomize rule.
3. Observe that lifeviewer miraculously pulls out one of the 262,144 OTL rules from the space of countless undecillions of INT rules.

This glitch may occur for other rulespaces as well, I haven't yet tested it.

Not sure whether I understand correctly what you mean by "OTL", the forum search returns no relevant results: search.php?keywords=OTL
The quote below may be relevant, though.

Personally I don't see lack of such a feature as a "glitch".
I think generating "random" isotropic 2-state cellular automata with range-1 Moore neighbourhood would be more confusing than helpful.

• The space is very large, and there are different subspaces. (B0, A0 B1c, A01c B1e, A01 B2a, A012a, ... ???) It is unclear whether to stay within a subspace (defined in some way or another), or to allow "jumping" between subspaces.
• Different birth/survival conditions have significantly different consequences, even within the same subspace (defined in some way or another). How to assign probabilities?

(These issues are already noticeable for Life-like cellular automata to some extent, but the space of Life-like cellular automata is sufficiently small, and I think in that case the usefulness of "give me a pseudorandom ruleset" outweights the problems with the idea / lack of an unique natural way to implement it.)

Things would become even more confusing with RuleLoader-compatible multistate cellular automata. Do you want to make the number of states pseudorandom, or you would prefer to randomize only the actual rules that are followed? Do you want to make the symmetries of the ruleset pseudorandom (i.e. switch between "none", "permute", "rotate4reflect")? Should the number of listed rules in a ruletable be kept constant, or it can change too? What to do with those cellular automata that are defined by a rule tree?

[...]

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).
[...]