Pattern viewer for forum threads

For discussion directly related to ConwayLife.com, such as requesting changes to how the forums or home page function.
User avatar
rowett
Moderator
Posts: 3776
Joined: January 31st, 2013, 2:34 am
Location: UK
Contact:

Re: Pattern viewer for forum threads

Post by rowett » December 5th, 2021, 4:00 am

muzik wrote:
December 1st, 2021, 10:12 am
The bounding box results on Identify aren't always consistent for hexagonal patterns when rotated, so I don't know if they should be kept in the output.
Hexagonal and triangular grids still use rectangular bounding boxes. It's a low priority item on the backlog to change this but if/when it happens it'll be fixed for Identify also.

User avatar
rowett
Moderator
Posts: 3776
Joined: January 31st, 2013, 2:34 am
Location: UK
Contact:

Re: Pattern viewer for forum threads

Post by rowett » December 5th, 2021, 4:01 am

muzik wrote:
December 4th, 2021, 12:22 pm
Can Identify be made to state the generation a pattern dies at in a message like is done in normal playback?
Yes, done.
muzik wrote:
December 4th, 2021, 12:22 pm
I also noticed that Select All still creates a 1x1 selection box even on empty patterns:
Select All now cancels any selection on empty patterns.
muzik wrote:
December 4th, 2021, 12:22 pm
In addition, Identifying the above empty pattern will advance it one generation and state that all cells died, but subsequent identifications correctly declare it an empty pattern (capitalized differently from the new pattern message).
Fixed, thanks.

User avatar
rowett
Moderator
Posts: 3776
Joined: January 31st, 2013, 2:34 am
Location: UK
Contact:

Re: Pattern viewer for forum threads

Post by rowett » December 5th, 2021, 9:45 am

muzik wrote:
November 30th, 2021, 11:19 pm
Is Select All supposed to include history states? This was fixed for outer-totalistic rules, but PCA still seems affected, but I don't know if they're handled differently.
Fixed, thanks.

hotdogPi
Posts: 1586
Joined: August 12th, 2020, 8:22 pm

Re: Pattern viewer for forum threads

Post by hotdogPi » December 5th, 2021, 10:06 am

One of your recent LifeViewer updates made it almost impossible to see my list of discoveries (User:hotdogPi/My discoveries), even after putting it in collapsed sections and opening it one at a time.
User:HotdogPi/My discoveries

Periods discovered: 5-16,⑱,⑳G,㉑G,㉒㉔㉕,㉗-㉛,㉜SG,㉞㉟㊱㊳㊵㊷㊹㊺㊽㊿,54G,55G,56,57G,60,62-66,68,70,73,74S,75,76S,80,84,88,90,96
100,02S,06,08,10,12,14G,16,17G,20,26G,28,38,47,48,54,56,72,74,80,92,96S
217,486,576

S: SKOP
G: gun

User avatar
rowett
Moderator
Posts: 3776
Joined: January 31st, 2013, 2:34 am
Location: UK
Contact:

Re: Pattern viewer for forum threads

Post by rowett » December 5th, 2021, 6:04 pm

hotdogPi wrote:
December 5th, 2021, 10:06 am
One of your recent LifeViewer updates made it almost impossible to see my list of discoveries (User:hotdogPi/My discoveries), even after putting it in collapsed sections and opening it one at a time.
Your discoveries page has 158 LifeViewers. You'll need to split it into multiple pages if you want to view it on a device with low memory (e.g. phone, tablet).
Collapsing sections does not change how much memory is used since the LifeViewers still exist. They are just not being displayed when the section is collapsed.

hotdogPi
Posts: 1586
Joined: August 12th, 2020, 8:22 pm

Re: Pattern viewer for forum threads

Post by hotdogPi » December 5th, 2021, 6:37 pm

rowett wrote:
December 5th, 2021, 6:04 pm
hotdogPi wrote:
December 5th, 2021, 10:06 am
One of your recent LifeViewer updates made it almost impossible to see my list of discoveries (User:hotdogPi/My discoveries), even after putting it in collapsed sections and opening it one at a time.
Your discoveries page has 158 LifeViewers. You'll need to split it into multiple pages if you want to view it on a device with low memory (e.g. phone, tablet).
Collapsing sections does not change how much memory is used since the LifeViewers still exist. They are just not being displayed when the section is collapsed.
I'm on my laptop. I can actually read the collapsed form just fine, but uncollapsing a single section takes an extremely long time and may not finish fully.
User:HotdogPi/My discoveries

Periods discovered: 5-16,⑱,⑳G,㉑G,㉒㉔㉕,㉗-㉛,㉜SG,㉞㉟㊱㊳㊵㊷㊹㊺㊽㊿,54G,55G,56,57G,60,62-66,68,70,73,74S,75,76S,80,84,88,90,96
100,02S,06,08,10,12,14G,16,17G,20,26G,28,38,47,48,54,56,72,74,80,92,96S
217,486,576

S: SKOP
G: gun

User avatar
muzik
Posts: 5612
Joined: January 28th, 2016, 2:47 pm
Location: Scotland

Re: Pattern viewer for forum threads

Post by muzik » December 6th, 2021, 12:02 am

Some annoyances regarding viewer thumbnails on touchscreen devices:

- If a thumbnail is tapped, but that tap is dragged and released in a region outside of the viewer thumbnail, the viewer window expands anyway. This behaviour is not reflected on desktop computers with the mouse when clicked and dragged.

Code: Select all

#C [[ THUMBNAIL ]]
x = 30, y = 3, rule = B3/S23
bo6bo12bo6bo$6bo16bo$3o3b2o14b2o3b3o!
- Selecting text from a text box and having the touched region enter a viewer thumbnail or expanded viewer will cancel the resizing of the selection and cause the viewer window to take over control (or, if the viewer window is already being controlled, for selections resizing to occur at the same time as something like zooming, which is weird):

Code: Select all

x = 10, y = 30, rule = B3/S23
2bob5o$2o2b4o$obo3b3o$bo2b2ob2o$bobob3obo$ob3o2b3o$3o$2bo4bobo$ob2o2b
3o$b3o3b2o$bo2bobobo$2b2ob3o$2b4o3bo$ob2o4bo$o2bobobobo$obobobo2bo$o4b
ob3o$3b5obo$3b2ob2obo$b3ob4o$2obo2b3o$o2b2ob4o$3obob3o$4o2b4o$b2ob3obo
$bob2ob2obo$3b2o2b3o$2bo2bob2o$2o2bob2o$2bobo2b3o!
To assist with debugging cases like that directly above as well as for general utility purposes, could some visual indicator be added to viewer windows/regions (such as a square in one of the corners which changes color accordingly) to state when the viewer is "in control/focus" or not (the state of "control" being what the viewer tells you to do with the "Click to control" message)?

Another random suggestion: button to shrink a selection to the bounding box of the contained cells (similarly to the option that does it automatically for all selections, just a manual version like Golly has).

User avatar
rowett
Moderator
Posts: 3776
Joined: January 31st, 2013, 2:34 am
Location: UK
Contact:

Re: Pattern viewer for forum threads

Post by rowett » December 6th, 2021, 2:26 am

hotdogPi wrote:
December 5th, 2021, 6:37 pm
I'm on my laptop. I can actually read the collapsed form just fine, but uncollapsing a single section takes an extremely long time and may not finish fully.
The advice stands: you'll need to break up that page. It's using too much memory on your laptop to work well.

User avatar
muzik
Posts: 5612
Joined: January 28th, 2016, 2:47 pm
Location: Scotland

Re: Pattern viewer for forum threads

Post by muzik » December 6th, 2021, 2:38 pm

KILLGLIDERS seems to work different for the following otherwise functionally identical patterns, and in the former case, replaces longevity cells with was-never-alive cells:

Code: Select all

x = 36, y = 9, rule = R1,C0,M0,S2..3,B3..3,NM
8b2o10bo$8b2o10bobo$20b2o7$3bo6bo$3bo2b2o2bo$2b10o$14o$2b10o9b2o$2b10o
9b2o$b12o$b12o$2b10o$2b10o$14o$2b10o$3bo2b2o2bo$3bo6bo!
[[ KILLGLIDERS THEME Red ]]

Code: Select all

x = 36, y = 9, rule = B3/S23
8b2o10bo$8b2o10bobo$20b2o7$3bo6bo$3bo2b2o2bo$2b10o$14o$2b10o9b2o$2b10o
9b2o$b12o$b12o$2b10o$2b10o$14o$2b10o$3bo2b2o2bo$3bo6bo!
[[ KILLGLIDERS THEME Red ]]

User avatar
muzik
Posts: 5612
Joined: January 28th, 2016, 2:47 pm
Location: Scotland

Re: Pattern viewer for forum threads

Post by muzik » December 6th, 2021, 3:01 pm

Despite yielding an empty pattern there is no "new pattern" message (or any error messages indicating what the box contains cannot be loaded) when viewing the following RLE:
c0b0p0 wrote:
May 23rd, 2015, 4:35 pm
The puffer is below. There is no known glider in this rule.

Code: Select all

x = 5, y = 17, rule = Phoenix
2.B$2.B$5B$B.B.B3$B.B.B$5B2$5B$B.B.B5$B.B.B$2.B!

User avatar
muzik
Posts: 5612
Joined: January 28th, 2016, 2:47 pm
Location: Scotland

Re: Pattern viewer for forum threads

Post by muzik » December 6th, 2021, 3:21 pm

Select All still seems to include dead cells for Generations patterns:

Code: Select all

x = 29, y = 9, rule = 23/3/3
21.2A4.2A$18.3A6.2A$15.3A$13.2A$A10.2A$A8.2A$A6.2A$.A3.2A$.4A!
[[ STARTFROM 4 ]]
However, this specifically only seems to affect those with more than 2 states, and only happens if not all cells are dead:

Code: Select all

x = 24, y = 24, rule = 23/3/2
22b2o$22b2o8$12b3o$14bo$13bo$9bo$9bobo$9b2o8$2o$2o!
[[ STARTFROM 30 ]]

Code: Select all

x = 23, y = 9, rule = 23/3/3
21.2A$18.3A$15.3A$13.2A$A10.2A$A8.2A$A6.2A$.A3.2A$.4A!
[[ STARTFROM 4 ]]
Also note how the first and third patterns here evolve differently in that STARTFROM makes the tick shape die completely but running it manually still has one dying cell in the fourth generation.

----

Another weird thing in the same vein - Select All on a pattern composed entirely of gray cells rejects said gray cells, but add any alive cell into the mix and they're completely welcomed:

Code: Select all

x = 9, y = 5, rule = LifeHistory
3F.F.F.F$F3.F.F.F$3F.F2.F$2.F.F.F.F$3F.F.F.F!

Code: Select all

x = 21, y = 5, rule = LifeHistory
3F.F.F.F.3A.3A.3A$F3.F.F.F.A.A.A.A.A$3F.F2.F2.A.A.A.A.3A$2.F.F.F.F.A.
A.A.A.A$3F.F.F.F.3A.A.A.3A!
Identify also lists the former as an empty pattern, which also seems off.

Yet another thing I've noticed while testing this that having Select enabled and tapping on the screen allows for an action to be undone even though no selection has truly happened.

----

One more thing: Identify used on a pattern that dies will cause the camera to zoom out for some reason. This doesn't make sense to me. Using HistoryFit and then running the pattern until its death also causes it to zoom out like this, and I have no idea why.

One cell test cases:

Code: Select all

x = 1, y = 1, rule = B3/S23
o!

Code: Select all

x = 1, y = 1, rule = B3/S23H
o!

Code: Select all

x = 1, y = 1, rule = B3/S23L
o!
Diehards:

Code: Select all

x = 8, y = 2, rule = B3/S23
bo4b2o$3o3bo!

Code: Select all

x = 49, y = 38, rule = B2/S34H
o$o$o$o$o$o$o$o$o$o$o$o$o$o$o$o$o$o$o$o$o$o$o$o$o$o$o33bo$o35bo$o$o32b
o5bo$o34bo3bo4bo$o43bo$o32bo4bo$33bo4bo4bo$43bo2$36bo11bo$36bo11bo!

Code: Select all

x = 7, y = 17, rule = B3/S23L
b3o$3o14$4b3o$3b3o!
Generations instead zooms in on the area where the last cell would have been (but is not due to the aforementioned bug):

Code: Select all

x = 23, y = 9, rule = 23/3/3
21.2A$18.3A$15.3A$13.2A$A10.2A$A8.2A$A6.2A$.A3.2A$.4A!

lemon41625
Posts: 344
Joined: January 24th, 2020, 7:39 am
Location: 小红点 (if you know where that is)

Re: Pattern viewer for forum threads

Post by lemon41625 » December 7th, 2021, 2:39 am

While I was coding a cellular automaton simulation library, I noticed that the heat calculated by LifeViewer for this ship is incorrect.

Code: Select all

x = 0, y = 0, rule = B02345/S0123
bo$o$3bo$2b2o$b2o$2o$2o$bo!
In the first generation, 5 cells die and 18 cells are born. So the heat for the 1st generation is 23.

However, LifeViewer lists the maximum heat as 18 which is lower than the heat in the 1st generation.
Download CAViewer: https://github.com/jedlimlx/Cellular-Automaton-Viewer

Supports:
BSFKL, Extended Generations, Regenerating Generations, Naive Rules, R1 Moore, R2 Cross and R2 Von Neumann INT
And some others...

User avatar
muzik
Posts: 5612
Joined: January 28th, 2016, 2:47 pm
Location: Scotland

Re: Pattern viewer for forum threads

Post by muzik » December 7th, 2021, 6:15 am

Selection options and transformations seem to work oddly for freshly-dead history cells in LifeHistory. For flips and rotations, history cells that remain history cells after the transformation will appear fully lit, whereas those that fill in was-never-alive regions appear long since dead. This is inconsistent with moving selections with the arrow buttons, in which case they all become fully dark.

Code: Select all

x = 28, y = 10, rule = LifeHistory
23.A2$6B14.6A$6B14.6A$6B12.A.6A$6B14.6A.A$6B14.6A$6B14.6A2$22.A!
Automatic selection shrinking also appears to work oddly on this pattern, for example selecting the dead square on the left before simulating anything (and then only), or on the right after all cells die in which only some of the cells are selected (in which case it seems to select the region that was last alive).

Random suggestion: can selections be made (optionally?) to display a center of the selection and/or two planes of symmetry to make it easy to tell if a rotation will take place around the center, edge or corner of a square?

User avatar
muzik
Posts: 5612
Joined: January 28th, 2016, 2:47 pm
Location: Scotland

Re: Pattern viewer for forum threads

Post by muzik » December 7th, 2021, 6:47 am

muzik wrote:
December 6th, 2021, 2:38 pm
KILLGLIDERS ...

Code: Select all

x = 36, y = 9, rule = R1,C0,M0,S2..3,B3..3,NM
8b2o10bo$8b2o10bobo$20b2o7$3bo6bo$3bo2b2o2bo$2b10o$14o$2b10o9b2o$2b10o
9b2o$b12o$b12o$2b10o$2b10o$14o$2b10o$3bo2b2o2bo$3bo6bo!
[[ KILLGLIDERS THEME Red ]]

Code: Select all

x = 36, y = 9, rule = B3/S23
8b2o10bo$8b2o10bobo$20b2o7$3bo6bo$3bo2b2o2bo$2b10o$14o$2b10o9b2o$2b10o
9b2o$b12o$b12o$2b10o$2b10o$14o$2b10o$3bo2b2o2bo$3bo6bo!
[[ KILLGLIDERS THEME Red ]]
I also noticed that at the time of death, the T menu (and Graph if you inspect it closely enough) reports the number of living cells in the first case as 11 even though it should be 8 as in the second.

Deleting the two blocks and then doing Select All also highlights a 1x1 area in the pattern. AutoFit also zooms in on this cell. The population remains at 11 despite these two blocks being removed as well.

Even filling the entire region with cells and clearing it still keeps the population at 11.


EDIT: Turns out the two blocks used to create a bounding box/diamond are completely irrelevant for general OT, so this pattern yields the same weird effects:

Code: Select all

x = 23, y = 23, rule = R1,C0,M0,S2..3,B3..3,NM
20bo$20bobo$20b2o7$3bo6bo$3bo2b2o2bo$2b10o$14o$2b10o$2b10o$b12o$b12o$
2b10o$2b10o$14o$2b10o$3bo2b2o2bo$3bo6bo!
[[ KILLGLIDERS THEME Red ]]

User avatar
rowett
Moderator
Posts: 3776
Joined: January 31st, 2013, 2:34 am
Location: UK
Contact:

Re: Pattern viewer for forum threads

Post by rowett » December 7th, 2021, 4:21 pm

muzik wrote:
December 6th, 2021, 12:02 am
If a thumbnail is tapped, but that tap is dragged and released in a region outside of the viewer thumbnail, the viewer window expands anyway. This behaviour is not reflected on desktop computers with the mouse when clicked and dragged.
This is fixed and will be in the next release.
muzik wrote:
December 6th, 2021, 12:02 am
Selecting text from a text box and having the touched region enter a viewer thumbnail or expanded viewer will cancel the resizing of the selection and cause the viewer window to take over control
This should also be fixed since I believe it's the same root cause as the issue above.
muzik wrote:
December 6th, 2021, 12:02 am
To assist with debugging cases like that directly above as well as for general utility purposes, could some visual indicator be added to viewer windows/regions (such as a square in one of the corners which changes color accordingly) to state when the viewer is "in control/focus" or not (the state of "control" being what the viewer tells you to do with the "Click to control" message)?
It's not helpful on touch devices since the window never actually gets focus.
muzik wrote:
December 6th, 2021, 12:02 am
Another random suggestion: button to shrink a selection to the bounding box of the contained cells (similarly to the option that does it automatically for all selections, just a manual version like Golly has).
I've put this on the backlog.

User avatar
rowett
Moderator
Posts: 3776
Joined: January 31st, 2013, 2:34 am
Location: UK
Contact:

Re: Pattern viewer for forum threads

Post by rowett » December 7th, 2021, 4:27 pm

muzik wrote:
December 6th, 2021, 2:38 pm
KILLGLIDERS seems to work different for the following otherwise functionally identical patterns
KILLGLIDERS is only supported for standard 2-state patterns. I've disabled it in the next build for other rules types (including 2-state HROT).

User avatar
rowett
Moderator
Posts: 3776
Joined: January 31st, 2013, 2:34 am
Location: UK
Contact:

Re: Pattern viewer for forum threads

Post by rowett » December 7th, 2021, 4:28 pm

muzik wrote:
December 6th, 2021, 3:01 pm
Despite yielding an empty pattern there is no "new pattern" message (or any error messages indicating what the box contains cannot be loaded) when viewing the following RLE
This is fixed and will be in the next released build.

User avatar
rowett
Moderator
Posts: 3776
Joined: January 31st, 2013, 2:34 am
Location: UK
Contact:

Re: Pattern viewer for forum threads

Post by rowett » December 7th, 2021, 4:44 pm

lemon41625 wrote:
December 7th, 2021, 2:39 am
While I was coding a cellular automaton simulation library, I noticed that the heat calculated by LifeViewer for this ship is incorrect.
This is because LifeViewer only looks at alternate generations for B0 rules which is enough to detect oscillator or spaceship period but will lead to incorrect Heat. It actually makes no sense to compute Heat in this case so I'll probably hide it for B0.

User avatar
muzik
Posts: 5612
Joined: January 28th, 2016, 2:47 pm
Location: Scotland

Re: Pattern viewer for forum threads

Post by muzik » December 8th, 2021, 11:31 am

Could the lower bound for MAXGRIDSIZE be lowered somewhat as to allow for easier experimentation with the boundaries, as well as (mainly for larger sizes) possibly an option to make the out of bounds area always appear by default rather than only when approached by living cells (or some option to set the distance a cell has to be from it to make it visible such that it can be set to scale appropriately depending on maxgridsize's definition)?

Code: Select all

x = 4, y = 9, rule = B3/S23
b2o$bo$bo$2o2$2b2o$2bo$2bo$b2o!
[[ MAXGRIDSIZE 11 ]]

Code: Select all

x = 4, y = 9, rule = B3/S23
b2o$bo$bo$2o2$2b2o$2bo$2bo$b2o!
[[ MAXGRIDSIZE 10 ]]

Code: Select all

x = 4, y = 9, rule = B3/S23
b2o$bo$bo$2o2$2b2o$2bo$2bo$b2o!
[[ MAXGRIDSIZE 9 ]]

Code: Select all

x = 4, y = 9, rule = B3/S23
b2o$bo$bo$2o2$2b2o$2bo$2bo$b2o!
[[ MAXGRIDSIZE 8 ]]

Code: Select all

x = 4, y = 9, rule = B3/S23
b2o$bo$bo$2o2$2b2o$2bo$2bo$b2o!
[[ MAXGRIDSIZE 7 ]]

Another suggestion: current selection is based on rectangles, meaning that Select All has a secondary purpose in making it easy to see the current bounding box of a pattern. Could some option be added for "von-Neumann" selections as well, which allow for 45-degree rectangular selections, which can therefore be used to display a pattern's bounding diamond instead (alongside its utility in selections where the normal rectangle might not be the best option)?

Code: Select all

x = 26, y = 7, rule = B3/S23
26o$o3b11o3b5ob2o$ob13obob5ob2o$ob3obobo2bo2bo3bo3bo2bo$ob3obobob2ob2o
b3obobob2o$o3bo3bob2ob2o3bobobo2bo$26o!

Code: Select all

x = 30, y = 30, rule = B3/S23
10bo$9b3o$8b5o$7b7o$6b9o$5b11o$4b13o$3b15o$2b2o3b12o$b3ob15o$4o3bobobo
3bo3b2o$b5obobob3ob3ob3o$2b2o3bo3bo3bo3b4o$3b21o$4b6o3bo3bob6o$5b5obob
ob3ob7o$6b4o3bo3bo2b7o$7b3ob5obob9o$8b2o3bo3bo2b9o$9b21o$10b6o3b3ob6o$
11b5obob3ob5o$12b4o3bo3b4o$13b3ob3obob3o$14b2o3bo3b2o$15b9o$16b7o$17b
5o$18b3o$19bo!
Another thing to note: "Click to control" is a common sight when using LifeViewer on touchscreen devices, which doesn't make sense as there's nothing to "click"

User avatar
rowett
Moderator
Posts: 3776
Joined: January 31st, 2013, 2:34 am
Location: UK
Contact:

Re: Pattern viewer for forum threads

Post by rowett » December 8th, 2021, 1:35 pm

muzik wrote:
December 8th, 2021, 11:31 am
Could the lower bound for MAXGRIDSIZE be lowered somewhat as to allow for easier experimentation with the boundaries, as well as (mainly for larger sizes) possibly an option to make the out of bounds area always appear by default rather than only when approached by living cells (or some option to set the distance a cell has to be from it to make it visible such that it can be set to scale appropriately depending on maxgridsize's definition)?
In a future build it will be possible to set MAXGRIDSIZE separately for width and height. The minimum width will stay at 9 (512 pixels) but the minimum height will reduce to 8 (256 pixels).
No changes are planned to the out of bounds rendering.

User avatar
muzik
Posts: 5612
Joined: January 28th, 2016, 2:47 pm
Location: Scotland

Re: Pattern viewer for forum threads

Post by muzik » December 8th, 2021, 2:45 pm

Can 2-state random fill be made to use the state currently selected for drawing, rather than only using the final state?

Code: Select all

#C Quick template for Select All
x = 20, y = 20, rule = Fredkin_mod6_vonNeumann
A19$19.A!
Also can the "none" rule have drawing and other operations enabled? This would make multistate editing capabilities easier to test things out with, as well as allow for other things (e.g. creating pixel art with the default Golly shades)

User avatar
muzik
Posts: 5612
Joined: January 28th, 2016, 2:47 pm
Location: Scotland

Re: Pattern viewer for forum threads

Post by muzik » December 8th, 2021, 3:35 pm

Patterns run using AUTOFIT and TRACK tend to "jolt" frequently, even though the evolution is always predictable. Is this something that can be fixed, or is it something that we have to put up with because we can't always have a solid 60 FPS all the time resulting in these frame glitches happening?

(In addition, TRACK also seems to "drift" in position even though by definition the amount it moves should be constant. Every time it jumps back, the cells still are aligned as they were directly beforehand. GRID highlights this phenomenon much more clearly.)

Code: Select all

x = 1, y = 1, rule = PCA_4
A!
[[ AUTOFIT ]]

Code: Select all

x = 1, y = 1, rule = MAPAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAD//////////////////////////////////////////w
o!
[[ AUTOFIT ]]

Code: Select all

x = 1, y = 1, rule = MAPAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAD//////////////////////////////////////////w
o!
[[ TRACK 1 1 ]]
I've tried having Frame Times enabled when running these and there may be a correlation between the two, but I can't tell entirely since I get what looks to be a solid 60fps either way.

EDIT: After having posted this they seem considerably more unstable - I can only assume this is due to the page being bigger and therefore meaning these have more difficulty catching up, causing a fluctuating FPS between 59 and 60.

EDIT2: This p2 example might make the oscillating case clearer:

Code: Select all

x = 1, y = 1, rule = MAPAP8A/wD/AP8A/wD/AP8A/wD/AP8A/wD/AP8A/wD/AP8A/wD/AP8A/wD/AP8A/wD/AP8A/wD/AP8A/wD/AP8A/w|MAPAAAAAP////8AAAAA/////wAAAAD/////AAAAAP////8AAAAA/////wAAAAD/////AAAAAP////8AAAAA/////w
o!
[[ AUTOFIT ]]
EDIT3: Further testing shows that this is not only also the case for patterns which cause the zoom to change, but also is probably responsible for the flickering of p2 periodic cells that's annoyed me for years now:

Code: Select all

x = 20, y = 1, rule = B0/SV
o5bo!
[[ AUTOFIT ]]
(The following pattern also causes a funny effect due to the zoom changes involved, however this is probably not a bug.)

Code: Select all

x = 20, y = 1, rule = B0/SV
o41bo!
[[ AUTOFIT ]]
Again, all of this is probably just not possible to fix and is either an intrinsic limitation or hardware-based.

----

I must say that I do kind of miss the "Build X is now live on the forums and wiki" posts and the summaries of what has been implemented since the previous build. However, since it's been over a year since the last one, I fear such a post would end up being impractically large...

User avatar
muzik
Posts: 5612
Joined: January 28th, 2016, 2:47 pm
Location: Scotland

Re: Pattern viewer for forum threads

Post by muzik » December 9th, 2021, 10:29 am

Is there any way to change the color of text which is cyan by default via scripts, red text from decoding errors, white text shown on actions such as "Play", the green highlight when changing sliders, the red flashing of the FPS box, as well as the color of snow?

There's also at least two cases where invalid patterns are still able to be shown in the viewer via forum code boxes: patterns which are too big, and patterns which contain a state outside of the supported range. (Using Select on the former has some weird effects: it still appears to be selecting stuff, but no box shows up.)

Code: Select all

x = 4, y = 1, rule = B3/S23
1111o!
[[ MAXGRIDSIZE 9 ]]

Code: Select all

x = 4, y = 4, rule = LifeHistory
4E$E2GE$E2GE$4E!
Personally I'm of the opinion that invalid patterns should always be able to be shown in the viewer provided that it's a correctly formatted RLE. Completely blocking them from being opened makes debugging more difficult than it should be, since there's no way to debug them (besides pasting them into the box on the other website, which is needlessly inconvenient, or using viewer tags instead of code tags, which is also not convenient) and figure out what the problem with the contained pattern is. The viewer is capable of showing error messages, so why should it be barred from doing so?

Code: Select all

x = 4, y = 4, rule = B3-j3j/S23
bo$o$o2bo$3o!

Code: Select all

x = 4, y = 4, rule = R1,C0,S2-3,B3History
bo$o$o2bo$3o!

Code: Select all

x = 10, y = 10, rule = R1,C0,S2-3,B3:T100,0
3o5b2o$o8bo$2bo4bobo$3b4o$3bo2bo$3bo2bo$3b4o$obo4bo$o8bo$2o5b3o!

Code: Select all

x = 9, y = 18, rule = B3/S23:S30*
b2o$bo$bo$2o$3b2o$3bo$3bo$3bo$2b2o$5b2o$5bo$5bo$5bo$4b2o$7b2o$7bo$7bo$
6b2o!

Code: Select all

x = 3, y = 3, rule = R0,C2,S0,B0
3o$3o$3o!

Code: Select all

x = 3, y = 3, rule = /2/1
3A$A.A$3A!

Code: Select all

x = 10, y = 3, rule = B3/S239
b2o4bo$obo4bobo$2bo4b2o!

Code: Select all

x = 8, y = 8, rule = R1,C0,S2-3,B0-3
b6o$2o4b2o$o2b2o2bo$ob4obo$ob4obo$o2b2o2bo$2o4b2o$b6o!

Code: Select all

x = 10, y = 10, rule = R1,C0,S2,B3,9
2o6b2o$ob6obo$b2o2bob2o$bo2b2o2bo$b6obo$bob6o$bo2b2o2bo$b2obo2b2o$ob6o
bo$2o6b2o!

Code: Select all

x = 8, y = 8, rule = 2/03/3
3bo$3bobo$bo$6b2o$2o$6bo$2bobo$4bo!

Code: Select all

x = 8, y = 8, rule = MAPERROR
o4bo!

Code: Select all

x = 8, y = 8, rule = W-1
o!

Code: Select all

x = 8, y = 8, rule = W1
o!

Code: Select all

x = 8, y = 8, rule = W256
o!

Code: Select all

x = 7, y = 3, rule = M14,15,13,12,11,10,9,8,7,6,5,4,3,2,1,0
o3b3o$3o2bo$bo!

Code: Select all

x = 7, y = 3, rule = M15,14,13,12,11,10,9,8,7,6,5,4,3,2,1,1
o3b3o$3o2bo$bo!
A discovery I made while setting up these error cases: aren't both of these patterns/rules equivalent? If so, should the top one not be rejected?

Code: Select all

x = 21, y = 20, rule = R1,C0,S0,12,B1,NL:T21,21
4b13o3bo$3bo7b4o3b3o$2bo7bo2b2o2bo2bo$2bo7bo2bob2o3bo$bo7bo3b2obo3bo$b
o6bo3bo4bo2bo$o6bo2b3o5b3o$o6bobo2bo3b5o$o5bobo3bob2o2b3o$o5b2o2bob2o
3bob2o$o4bo5b2o3bo3bo$o3bo5b2o3bo4bo$o2bo5bobo2bo4b2o$obo4b2o2bobo5b2o
$b6o4bobo5b2o$11b2o6b2o$11b2o5bobo$11b2o5bobo$11b2o3b2o2bo$10b11o!

Code: Select all

x = 21, y = 20, rule = B1/S0ZL:T21,21
4b13o3bo$3bo7b4o3b3o$2bo7bo2b2o2bo2bo$2bo7bo2bob2o3bo$bo7bo3b2obo3bo$b
o6bo3bo4bo2bo$o6bo2b3o5b3o$o6bobo2bo3b5o$o5bobo3bob2o2b3o$o5b2o2bob2o
3bob2o$o4bo5b2o3bo3bo$o3bo5b2o3bo4bo$o2bo5bobo2bo4b2o$obo4b2o2bobo5b2o
$b6o4bobo5b2o$11b2o6b2o$11b2o5bobo$11b2o5bobo$11b2o3b2o2bo$10b11o!
EDIT: It seem that even valid triangular bounded grids are flawed. Shouldn't this cell turn on at T 32, keeping this arrangement symmetric?

Code: Select all

x = 2, y = 2, rule = B12/SL:T30,30
2o$2o!
[[ STOP 32 T 16 X 0 Y 0 T 32 X -15 Y -15 ZOOM 16 ]]
Last edited by muzik on December 9th, 2021, 12:00 pm, edited 1 time in total.

User avatar
muzik
Posts: 5612
Joined: January 28th, 2016, 2:47 pm
Location: Scotland

Re: Pattern viewer for forum threads

Post by muzik » December 9th, 2021, 11:46 am

Changing the graph's color may be responsible for a particularly nasty crash:

Code: Select all

x = 33, y = 27, rule = LifeHistory
obo!
[[ COLOR GRAPH 240 255 255 ]]

User avatar
rowett
Moderator
Posts: 3776
Joined: January 31st, 2013, 2:34 am
Location: UK
Contact:

Re: Pattern viewer for forum threads

Post by rowett » December 9th, 2021, 3:04 pm

muzik wrote:
December 9th, 2021, 11:46 am
Changing the graph's color may be responsible for a particularly nasty crash:

Code: Select all

x = 33, y = 27, rule = LifeHistory
obo!
[[ COLOR GRAPH 240 255 255 ]]
This is fixed and will be in the next release. Thanks for reporting!

Note that "GRAPH" is not a valid colour element.

Post Reply