Opened 2 years ago
Last modified 15 months ago
#6562 new defect
Atlas crashes on Linux (Fedora 36 using RPM latest version)
Reported by: | Dworv | Owned by: | |
---|---|---|---|
Priority: | Must Have | Milestone: | Alpha 27 |
Component: | Atlas editor | Keywords: | RPM Atlas |
Cc: | Dworv | Patch: |
Description
The first time I built a map using atlas, everything went great. However, I recently tried building a map and when sculpting, the mouse seemed to get stuck and it just kept expanding in one spot. For some reason, my cursor lost the ability to press any buttons at all. I had to use some tricky terminal magic to quit the process, as I couldn't press any buttons to use the system manager! Please look into this, I love making maps but this just makes it impossible.
Using the latest version according to RPM (XXV Yauna)
Change History (15)
follow-up: 2 comment:1 by , 2 years ago
comment:3 by , 2 years ago
Milestone: | → Alpha 26 |
---|
Adding a milestone so that it can be investigated and not forgotten.
[21:53:10] elexis #6562 doesnt have a milestone. it might be invalid (as in fix your OS), or it might be an RB and annoying to reproduce inside a VM
comment:4 by , 2 years ago
Replying to Dworv:
Replying to stanislas69:
How would I go about doing that?
Install gdb
, then run: gdb pyrogenesis
.
Could you check whether the same thing happens when only sculpting small bits at a time? Click shortly, wait until it finishes, repeat.
comment:5 by , 22 months ago
I run Fedora and can try to reproduce the bug if that helps. Where can I get the pre-release RPMs?
comment:6 by , 22 months ago
There are no pre-release RPMs. This seem to happen with the actual release. Any help would be appreciated.
follow-up: 13 comment:7 by , 22 months ago
Couldn't reproduce, even with a map full with random things on it.
@Dworv, is the behavior reproducable? If so, can you give me steps to do so, or share the map you were drawing?
comment:8 by , 22 months ago
I think it should be reproducable, since I reinstalled and it still happened from a blank slate. Just make a giant brush and rise the terrain, after a while of holding it should get stuck.
comment:10 by , 22 months ago
Here is what I did:
gdb pyrogenesis [...] run
In the opening 0ad window I picked the scenario editor. I put the new scenario editor on fullscreen, put the brush size to 100, the strength also to 100, and started raising. My machine was screaming because of the tiny intel graphics card, but apart from that I had no trouble with clicks or keyboard input not being recognized properly.
My system specs:
❯ neofetch .',;::::;,'. liquidat@russell .';:cccccccccccc:;,. ---------------- .;cccccccccccccccccccccc;. OS: Fedora release 36 (Thirty Six) x86_64 .:cccccccccccccccccccccccccc:. Host: 20W00050GE ThinkPad T14 Gen 2i .;ccccccccccccc;.:dddl:.;ccccccc;. Kernel: 5.18.10-200.fc36.x86_64 .:ccccccccccccc;OWMKOOXMWd;ccccccc:. Uptime: 2 hours, 14 mins .:ccccccccccccc;KMMc;cc;xMMc:ccccccc:. Packages: 2185 (rpm), 63 (flatpak) ,cccccccccccccc;MMM.;cc;;WW::cccccccc, Shell: nu 0.64.0 :cccccccccccccc;MMM.;cccccccccccccccc: Resolution: 1920x1080 :ccccccc;oxOOOo;MMM0OOk.;cccccccccccc: DE: GNOME 42.3.1 cccccc:0MMKxdd:;MMMkddc.;cccccccccccc; WM: Mutter ccccc:XM0';cccc;MMM.;cccccccccccccccc' WM Theme: Adwaita ccccc;MMo;ccccc;MMW.;ccccccccccccccc; Theme: Adwaita [GTK2/3] ccccc;0MNc.ccc.xMMd:ccccccccccccccc; Icons: Adwaita [GTK2/3] cccccc;dNMWXXXWM0::cccccccccccccc:, Terminal: tilix cccccccc;.:odl:.;cccccccccccccc:,. CPU: 11th Gen Intel i7-1165G7 (8) @ 4.700GHz :cccccccccccccccccccccccccccc:'. GPU: Intel TigerLake-LP GT2 [Iris Xe Graphics] .:cccccccccccccccccccccc:;,.. Memory: 9130MiB / 31721MiB '::cccccccccccccc::;,.
@dworv: Since this is repeatable for you, can you please share debuginfo when this happens? Also, what are your system specs?
comment:11 by , 21 months ago
Milestone: | Alpha 26 → Alpha 27 |
---|
comment:12 by , 21 months ago
I managed to run gdb on it, but there wasn't any helpful output. This was all that came from the logs
GAME STARTED, ALL INIT COMPLETE
[Thread 0x7fff6742a640 (LWP 102583) exited]
(atlas:102080): Gtk-CRITICAL **: 16:34:51.530: gtk_box_gadget_distribute: assertion 'size >= 0' failed in GtkScrollbar
After this, I began using a giant brush. Nothing appeared in the logs, when I held down on a single spot for about 5 seconds, I let go and took my cursor off. It kept trying to fill it. My suspicion is that the game accumulated a lot of requests to raise the area, but wasn't fast enough to complete them. The time between updates on the ground raising goes from half a second to around 5 seconds before just stopping. I don't know if it gets more expensive to raise the ground as it gets higher up, but after the point where the ground stops raising, my fans continue to roar and I have to kill -9
the process from the terminal, as my mouse buttons and scroll wheel stop working anywhere on the screen.
comment:13 by , 21 months ago
Replying to liquidat:
Couldn't reproduce, even with a map full with random things on it.
@Dworv, is the behavior reproducable? If so, can you give me steps to do so, or share the map you were drawing?
Its an empty map.
comment:14 by , 21 months ago
Here is my neofetch:
dworv@epicness
OS: Fedora Linux 36 (Workstation Edition
Host: MS-7C02 1.0
Kernel: 5.18.15-200.fc36.x86_64
Uptime: 18 hours, 30 mins
Packages: 2563 (rpm), 22 (flatpak)
Shell: bash 5.1.16
Resolution: 1920x1080, 1080x1920
DE: GNOME 42.3.1
WM: Mutter
WM Theme: Adwaita
Theme: Adwaita-dark [GTK2/3]
Icons: Numix-Circle [GTK2/3]
Terminal: gnome-terminal
CPU: AMD Ryzen 5 3600X (12) @ 3.800GHz
GPU: NVIDIA GeForce GTX 1650 SUPER
Memory: 4947MiB / 32041MiB
comment:15 by , 15 months ago
I have encountered the same issue when creating my first map. These crashes occur more often when the terrain is already strongly modified. Had a huge map with many trees and much terrain sculpting. My impression also was that this occurred more often in the edge of the map, but that may aswell be subjective.
Once Atlas froze, i just clicked the window until it grayed out and kde asked to kill unresponsive process pyrogenesis.
It looked like atlas was unable to halt the last sculpting operation, so that the terrain would just go higher and higher (simulation still moving), but no input was accepted, as described here.
I think this is not related to the unit amount, but maybe rather to the amount of terrain modifications which has been done. My map had huge setting and some higher terrain "layers".
System:
Arch linux
Intel(R) Core(TM) i5-4690 CPU @ 3.50GHz
GeForce GTX 970
16 Gib RAM
K Desktop Environment
Hey could you run the game through GDB? It looks like a performance issue.