Indie game storeFree gamesFun gamesHorror games
Game developmentAssetsComics
SalesBundles
Jobs
Tags

Smith and Winston

Explore and Destroy in this fun adventure, with twin stick combat and full level destruction · By Execution Unit, Charlie, MetalDazza

Linux issues

A topic by MetalDazza created Oct 12, 2018 Views: 1,157 Replies: 21
Viewing posts 1 to 8
Developer (1 edit)

A few users have reported issues on openSUSE and Arch Linux. 

The game doesn't start from the itch app:

this is most likely due to libsndio not being installed or version 7 being installed. The app is looking for libsndio.so.6.1 so installing libsndio and symlinking libsndio.so.6.1 to libsndio.so.7 should fix that.

next is that the app is looking for libcurl-gnutls, installing and symlinking fixes this. The next release won't use libcurl anyway so this issue should be fixed soon (it's part of a debug feature that is due to be removed).

lastly - 64bit inodes. The app is 64 bit but for some reason there is still a crash when running from a 64 bit file system. A work around is to move the app to a 32bit files system. We'll look into why this is happening and  fix ASAP.

Thanks for the bug reports and for your patience.

dazza

Developer

The libsndio and libcurl-gnutls issues should be fixed in the latest release now. Let me know if you come across anything else.

(1 edit)

Just thought I would give some feedback, so far the game seems to be running fine on my Archlinux system with a Fury running the opensource drivers. I've only gotten through I think to level 2, the level after the ship, and partway into the next level. So far it seems to be a solid 60 fps running in 4k, and it seems to handle it fairly well, the particle effects seem to run a bit on the heavy side, in the area near the ship, as when power limiting my Fury to 100 watts to help to reduce some of the heat it will drop down to around 35-40 fps, but overall seems to run well. It handles changes to the window well it seems, no issues with resizing the window. One thing I did notice is that it seems like I can't turn off VSync, it remains capped at 60 fps no matter what the toggle is set to. Also I'm not 100% sure if the shadow settings all work or not, I notice with low and medium they are visible, but it seems like it's reversed or something, notice the shadows more with Low and the higher I go they get more soft or something, but maybe it's just them being more accurate.

Also, I tried it with my Steam controller running through the sc-controller app as an xbox 360 controller and it seemed to detect it fine and be playable. Though one thing I might recommend is maybe making it possible for both mouse and joystick can work together, to make full use of the touchpads that it can have.

Also, thought I would let you know that the game segfaults on quitting with this backtrace:

Thread 1 "SmithAndWinston" received signal SIGSEGV, Segmentation fault.
0x00007ffff4d6680b in _int_free () from /usr/lib/libc.so.6
(gdb) bt
#0  0x00007ffff4d6680b in _int_free () from /usr/lib/libc.so.6
#1  0x00007ffff541a935 in _XFreeDisplayStructure () from /usr/lib/libX11.so.6
#2  0x00007ffff5408422 in XCloseDisplay () from /usr/lib/libX11.so.6
#3  0x00007ffff7d71382 in X11_DeleteDevice (device=0x555555f98710) at /home/dazza/Downloads/SDL2-2.0.8/src/video/x11/SDL_x11video.c:116
#4  0x00007ffff7d3d5f6 in SDL_VideoQuit_REAL () at /home/dazza/Downloads/SDL2-2.0.8/src/video/SDL_video.c:2771
#5  0x00007ffff7c89164 in SDL_QuitSubSystem_REAL (flags=29233) at /home/dazza/Downloads/SDL2-2.0.8/src/SDL.c:294
#6  SDL_Quit_REAL () at /home/dazza/Downloads/SDL2-2.0.8/src/SDL.c:354
#7  0x00005555556d256e in main ()

Will let you know of further findings.


Developer

Great feedback, thanks,

yeah the particles around the ship are REALLY heavy, bloody designers ;) OpenGL really doesn't like it and DX11 only just copes. We'll have to look at it. I'll look in to the vsync, that should work. Yeah, the shadows are odd, the higher quality they are the more blocky they are.... because they are more voxel-like, bit odd :)

I've got a reproduction of that crash and I think I know what it is. The progress and preferences are saved before it does it so it's not too fatal. I'll fix it though.

Thanks for helping to make the game better.

Developer

Had a look at those particles clustered around the ship and found they where doing a couple of unnecessary sorting passes. So I've fixed that and reduced their number by about 30%. Looks pretty much the same but should run better, thanks again for spotting this. (this will be in for the next release)

(+1)

Hi,

I recently bought your (early-access) game and sadly I have to file a bug report after having only played to the second portal (the one which should make you progress in game). The game gets stuck at the loading screen though, with following terminal output:

INFO: BSEd 2.8.3.4.5.4.88d9b6a
INFO: Voxity 4.5.4.88d9b6a
INFO: num displays 0
INFO: System Pref path - /home/stephan/.local/share/ExecutionUnit/SmithAndWinston/
INFO: Resource path - /home/stephan/.config/itch/apps/smith-and-winston/x86_64/
ERROR: GLOBALS are locked -- gSceneStats must be declared local or prefix with '__' for globals.    2
stack traceback:
    Scripts/utility/utility.lua:248: in function '__newindex'
    Scripts/vox/snw/hud.lua:3117: in function <Scripts/vox/snw/hud.lua:3099>
    Scripts/vox/snw/hud.lua:5054: in function '_select_menu_option'
    Scripts/vox/snw/hud.lua:5398: in function <Scripts/vox/snw/hud.lua:5378>
    Scripts/vox/snw/hud.lua:5300: in function '_update_input'
    Scripts/vox/snw/hud.lua:4499: in function <Scripts/vox/snw/hud.lua:4488>
ERROR: end traceback

I'm gaming on Fedora with amdgpu and mesa (mostly latest git snapshots). Latest version of the game which is 15 from Jan 14th. Maybe this is helpful information.
Developer(+1)

Hi nervenbuendel,

Sorry it's not running for you. I've just reinstalled the latest version on Ubuntu and I'm not seeing this. Can you run the game from the terminal with the parameter -console:

/home/stephan/.config/itch/apps/smith-and-winston/SmithAndWinston -console

this should make the app print a lot more information. If you could paste the output here or email me at dazza@executionunit.com whichever is easiest for you.

Thanks for your help.

Developer(+1)

Hi nervenbuendel,

Looking at the information you provided last night it looks like you were pressing "Continue Arcade Game" in the main menu? If that's the case can you send me your preferences file please? You should find it here:

/home/stephan/.local/share/ExecutionUnit/SmithAndWinston/userprefs

If you could email it to me at dazza@executionunit.com I'll see what's up. I don't think this is Linux specific.

(1 edit) (+1)

Email sent. :-)

And yes, I pressed continue arcade game.

I'm sorry but I've run into an error again: This is what stderr gives:

$ ./SmithAndWinston  

INFO: BSEd 2.8.4.4.6.3.c16df17 INFO: Voxity 4.6.3.c16df17 INFO: num displays 0 INFO: System Pref path - /home/stephan/.local/share/ExecutionUnit/SmithAndWinston/ INFO: Resource path - /home/stephan/.config/itch/apps/smith-and-winston/x86_64/ SmithAndWinston: ../src/gallium/auxiliary/util/u_blitter.c:2062: util_blitter_generate_mipmap: Assertion `!util_format_has_stencil(desc)' failed. ./SmithAndWinston: Zeile 21:  3789 Aborted                 (Speicherabzug geschrieben) ./x86_64/${EXENAME} ./data $@
Developer(+1)

Hi nervenduendal,

Could this be a Mesa 18 issue? has mesa changed recently? What version are you running.


(sorry for the late reply, I didn't get an email from itch).

I'm using Mesa 19.1-git snapshots. I'll try again after the next update and inform you about my results.

Developer(+1)

thanks, I've checked through the Mesa bug database but I can't find any bug like this for 19.1, only 18.1.

Sadly the problem persists after 2 more Mesa snapshots.

Developer(+1)

We're working toward a new release in the next few days and after that I'll dig a little deeper and see if there is any work around I can put in the engine. The new release does include an updated BGFX (graphics) library so it might just magically be fixed but I'll do some more digging. Thanks for being patient with us.

Developer(+1)

I'm getting this in Ubuntu as well when I install the open source drivers for my Radeon card. Two things that worry me is that the 19.1 .0 release of Mesa is flagged as a "development release" and package maintainers seem to be building it without NDEBUG defined (so asserts can be triggered). I'm not sure there is anything I can so about this I'm afraid, can you roll back to the previous version as the release notes mention? https://www.mesa3d.org/relnotes/19.1.0.html

I've reverted back to Mesa 18.3.3 and LLVM 7.0.1 but the problem still persists.

Developer(+1)

oh dear, this is a proper mess isn't it? I'm sorry it's not working for you. My Xubuntu is on Mesa 19.1.0-devel, LLVM 7.0.1 and I'm getting the assert as well. Annoyingly my Radeon R9 380 isn't fully supported by Gallium3D and I only get a few limited resolutions which is making this hard to debug (I usually use the amd-gpu proprietary drivers). I'll do some more digging today and maybe get in contact with the BGFX developers to see if they have an hints on a work around.

Thanks for your patience with this. Do you happen to remember which version of  Mesa etc. you used when it was working?

Developer(+1)

I've reported a bug to FreeDesktop.org, hopefully this will be resolved soon: https://bugs.freedesktop.org/show_bug.cgi?id=109754 I'm nit 100% sure if it's a radeon_dri bug or a packaging bug but we'll get to the bottom of it. Thanks for your patience.

(+1)

I guess the last working version was Mesa 19.0-git, but I'm not too sure about it. Maybe the cause for the error got backported to Mesa-stable.

Your graphics card (Volcanic Islands aka GCN 3) should be well supported by amdgpu if I recall correctly, just as mine.

Thanks for your efforts.