Golly bugs

[Katrina]

Golly really likes to save 50+MB temporary pattern files whenever I run an infinite growth simulation for long enough. Is there a way to prevent this from happening so I don't get slowdowns?

[Andrew]

Golly really likes to save 50+MB temporary pattern files whenever I run an infinite growth simulation for long enough. Is there a way to prevent this from happening so I don't get slowdowns?

For the particular rule in your example, use the HashLife algorithm rather than QuickLife. QuickLife saves temporary patterns in .rle files whereas HashLife saves .mc files which are typically much smaller and faster to write.

If you don't want Golly to save *any* temporary patterns then tick the "Disable Undo/Redo" option in the Edit menu. It's then your responsibility to save any patterns you don't want to lose.

[b-engine]

I get this error whenever I try to open Golly 5.0, even on Xorg.

Code: Select all

./golly: error while loading shared libraries: libgtk-x11-2.0.so.0: cannot open shared object file: No such file or directory

[Andrew]

I get this error whenever I try to open Golly 5.0, even on Xorg. ...

Sorry, but it seems very difficult (impossible?) to create a binary that will run on all flavors of Linux. You'll have to download the source code and follow the instructions in docs/Build.html. EDIT: Or maybe just install that missing library.

[LaundryPizza03]

On macOS 15.6.1 with Apple M2, Golly occasionally freezes while opening the menu bar.

I'm pretty sure this is a bug in macOS. I've seen the same issue happen occasionally in Terminal.app. Never seen it happen in Golly so far. Let me know if you manage to find steps to reproduce the problem in Golly.

In v4.3, it sometimes happens while attempting to open the rule dialog.

EDIT: Also v 5.0 under similar circumstances. Is macOS 26 known to experience this issue?

[Andrew]

EDIT: Also v 5.0 under similar circumstances. Is macOS 26 known to experience this issue?

What is macOS 26?

[b-engine]

What is macOS 26?

The newest version of macOS, and Apple changed the version to sync with the year.
The entire UI switched to "Liquid Glass" design language, but we won't have Liquid Glass on Golly anyway.

[Mr. T]

Linux-User here. I want suggest, that Golly only in X11 works (on my computer). All my tries with Wayland shipwrecked.I am able to open the load-window ONE time, Then all of Golly freezes. For Information: I use NixOS. Golly starts in Arch-Linux using Distrobox. What went wrong? Any Idea?

[Andrew]

Linux-User here. I want suggest, that Golly only in X11 works (on my computer).

Sorry, but we dropped support for X11 many years ago (too many bugs and limitations in the wxWidgets X11 libs). Golly 1.4 was the last release that had an X11 build (available on sourceforge if you wanted to try it).

To run (or build) Golly on Linux you need GTK2 installed. I suspect to get it running on NixOS you'll need to download the source tar ball and follow the instructions in docs/Build.html.

[Mr. T]

Linux-User here. I want suggest, that Golly only in X11 works (on my computer).

Sorry, but we dropped support for X11 many years ago (too many bugs and limitations in the wxWidgets X11 libs). Golly 1.4 was the last release that had an X11 build (available on sourceforge if you wanted to try it).

To run (or build) Golly on Linux you need GTK2 installed. I suspect to get it running on NixOS you'll need to download the source tar ball and follow the instructions in docs/Build.html.

The Problem lies at NixOS, I think. I tried COSMIC and KDE Plasma in its Wayland-form. Both fail! Only Plasma X11 works. Problem: NixOS has Wayland as default. To smuggle X11 in, you have to define the X-Server in detail. In one of its X11-options NixOS must present the correct gtk-library. Nor gtk as package nor Xwayland work.

[Aleph]

To run (or build) Golly on Linux you need GTK2 installed. I suspect to get it running on NixOS you'll need to download the source tar ball and follow the instructions in docs/Build.html.

Is there any plan to support GTK3? I believe that a lot of GNU+Linux distributions use GTK3 instead of GTK2 these days.

[Andrew]

Is there any plan to support GTK3?

It should be possible to build a GTK3 version of Golly using the current source code, but I wasn't able to get it working for some reason. I first installed the GTK3 headers and libraries via "sudo apt install libgtk-3-dev". Then I rebuilt the wxWidgets libs using these commands:

Code: Select all

cd ~/Downloads/wxWidgets-3.2.9
mkdir build-gtk
cd build-gtk
../configure --with-gtk=3 --disable-shared --enable-unicode --with-opengl
make
sudo make install
sudo ldconfig

Then did a cd to Golly's gui-wx folder, followed by a "make clean" and make. Everything seemed to compile and link ok but when I tried to run golly it failed to start up. All I got was a "Terminated" message. I don't know enough about Linux to debug what has gone wrong. Maybe my Mint 20 system is too old? I'd be interested to see if you or any other Linux users can use the above steps to successfully build a GTK3 version.

[qqd]

While experimenting with rule files, I stumbled upon this minor but weird bug when defining a variable with a comma (which would normally be considered a syntax error) in a rule file for Golly's RuleLoader algorithm to use (note that this has only been tested on 64-bit Windows 11):
Firstly, copy the following code and paste it in Golly (alternatively you can use 'Run Clipboard'):

Code: Select all

@RULE OneD4stateAdjustableRakesBugDemo

This is a 4-state 1D rule having adjustable rakes using 3 on states in a sea of off states
It may be possible to modify the rule to make it more interesting without changing the number of states

state 0 - black  - off
state 1 - yellow - still life
state 2 - red    - left moving photon
state 3 - green  - right moving photon

@TABLE
n_states:4
neighborhood:oneDimensional
symmetries:none
var any,0={0}
any,0,0,2,2
0,1,2,3
0,2,2,3
0,3,0,3
0,3,1,3
0,3,2,1
1,2,0,0
1,0,3,0
1,0,1,2
1,1,0,3
1,3,0,3
2,0,0,0
2,0,3,0
2,0,1,3
2,2,0,1
2,2,2,0
2,3,0,3
3,0,0,0
3,1,1,2
3,2,0,1
3,3,0,0
3,0,1,2
3,0,2,0
3,0,3,1

@NAMES

0 empty
1 dot
2 left photon
3 right photon

@COLORS

0   0   0   0
1 255 255   0
2 255   0   0
3   0 255   0

I was unable to isolate this bug in spite of several attempts to do so, so I am posting the full rule table that I modified rather than a simplified version. Strangely, Golly does not throw an error that the rule file is invalid for the above table with "any,0", but does throw an error for the simplified rules I attempted to create.
Second, paste this sample pattern to Golly and run the pattern in the new rule above (WARNING: FLASHING LIGHTS):

Code: Select all

x = 1, y = 1, rule = OneD4stateAdjustableRakesBugDemo
A!

The rule defined above uses a oneDimensional neighborhood. Yet it explodes in both the horizontal and vertical directions. Moreover, the rate at which it explodes increases in an exponential manner.
Since this has only occurred in the rule mentioned above, it might not be a very concerning bug, but it is worth investigating.

[Andrew]

While experimenting with rule files, I stumbled upon this minor but weird bug when defining a variable with a comma ...

I've fixed that particular bug for the next release, and improved the error reporting. It's still possible to make syntax errors that won't be detected but I'm reluctant to make too many changes to the parsing code (not written by me). Thanks for the report.

[SuperSupermario24]

The rule defined above uses a oneDimensional neighborhood. Yet it explodes in both the horizontal and vertical directions. Moreover, the rate at which it explodes increases in an exponential manner.
Since this has only occurred in the rule mentioned above, it might not be a very concerning bug, but it is worth investigating.

I think this is just how any rule table works if it has a B0 transition, they're not really supported very well.

[hotdogPi]

The rule defined above uses a oneDimensional neighborhood. Yet it explodes in both the horizontal and vertical directions. Moreover, the rate at which it explodes increases in an exponential manner.
Since this has only occurred in the rule mentioned above, it might not be a very concerning bug, but it is worth investigating.

I think this is just how any rule table works if it has a B0 transition, they're not really supported very well.

No legitimate rule should have the bounding box increase more than linearly (per direction) or quadratically (total). Exponential growth sounds like a HashLife error.

[Katrina]

Headerless extended RLEs (like of the form 9.3A2.$9.3A2.$8.4A2.$A7.B5A$A6.A2CBA2.$7.AC3A2.$6.3A.2A2.$6.BADBA3.$8.A5.!) do not get pasted properly, instead treated like a 1-dimensional bitmap with dots as 0s and everything else as 1.

Crossposted from Discord.

[Katrina]

The rule "R1,C3,S2-5,B6,8,NW111101111,110" and similar rules are completely broken, acting as if they were somehow B0 Generations rules.
All the cells within the bounding box of the pattern within the unbounded universe are flipped when running a pattern in the rule, as well. It does behave somewhat more normally in bounded universes, but still with strobing.

[I6_I6]

600th post!!
This is also a problem in LifeViewer.
In @TABLE in a .rule file, if I name a variable "var", and use it at the start of a transition line, it doesn't behave as expected.
For example, this rule should be B2/S012345678:

Code: Select all

x = 0, y = 0, rule = test
!
@RULE test
@TABLE
n_states:2
neighborhood:Moore
symmetries:permute
var var={0,1}
var,1,1,0,0,0,0,0,0,1

But it behaves like B/S012345678 instead.
Simply renaming the variable to something else fixes this:

Code: Select all

x = 0, y = 0, rule = test
!
@RULE test
@TABLE
n_states:2
neighborhood:Moore
symmetries:permute
var a={0,1}
a,1,1,0,0,0,0,0,0,1

I think it's because Golly treats any line in @TABLE that starts with "var" as a variable definition line. I don't know why anyone would name a variable "var", but I think it should still be fixed.

[rowett]

The rule "R1,C3,S2-5,B6,8,NW111101111,110" and similar rules are completely broken, acting as if they were somehow B0 Generations rules.

Thanks for the report.

The algo shouldn't allow a non-zero dead state weight (i.e. the first digit of ,110) since this leads to B0-type behaviour which is not supported.

I'll fix it in due course.

[Mr. T]

Linux

Problems with Wayland still exist. This time the system is: Cachy Os with plasma + Wayland. Same crab as with NixOS (Plasma and Cosmic). If you want load a pattern into Golly, Golly freezes and you can get it from the system only with some form of kill.

Thence here my appeal. Test your Linux-system and submit the result. Perhaps the bug is solvable, whether in Golly or Wayland or it is a me-problen.

Thank you for your cooperation.

[Antonin Duda]

I don't know if this is the fault of golly, or python itself, but file paths containing characters not within UTF-8 aren't read properly.
In case you're wondering, the folder with the question mark character is named "Antonín", which my local account name my laptop vendor set up.

[hotdogPi]

In hex rules, LifeHistory gray cells can still kill cells that are adjacent in Moore but not hexagonal.

[aengmore]

Seems Windows Golly cannot run properly in Winlator (a Windows emulator on Android). A dialog box says “Failed to create a temporary file name (error 267: The directory name is invalid.)” and the display area is completely black, rendering no grid, no cells, no graphic.

Winlator: https://github.com/brunodev85/winlator

Android Golly lacks so many functions compared with the desktop version So I try to use the Windows version on my phone, and it failed

Failed to create a temporary file name (error 267: 目录名无效。)

Screenshot_2026-06-30-10-07-55-65_b6022cb7d1a61f21e4fe914c8728a54c.jpg (169.42 KiB) Viewed 215 times

[NNlk05]

*snip*
Android Golly lacks so many functions compared with the desktop version So I try to use the Windows version on my phone, and it failed
*snip*

Since Android is a version of Linux, maybe running the Linux version will be smoother.