Pioneer crash at start #4174

Tichy · 2017-10-01T16:52:53Z

After updating my local git repo, I compiled Pioneer but the new version crash as soon as it starts.
Visually I just see the window appear then disappear.

Launching from the terminal I obtain this:

ver git (8e90cc7f9) on: Linux

System Name: Linux
Host Name: ono-sendai
Release(Kernel) Version: 4.13.3-1-ARCH
Kernel Build Timestamp: #1 SMP PREEMPT Thu Sep 21 20:33:16 CEST 2017
Machine Arch: x86_64
Domain Name: (none)

SDL Version 2.0.6
Initialized OpenGL 3.1, with extensions, renderer
invalid axis binding '' in config file for BindAxisPitch
invalid axis binding '' in config file for BindAxisRoll
invalid axis binding '' in config file for BindAxisYaw
started 7 worker threads
ShipType::Init()
Lua::Init()
Fonts:
font orbiteer:
- DejaVuSans.ttf 1.200000
- wqy-microhei.ttc 1.000000
- Orbiteer-Bold.ttf 1.000000
font pionillium:
- DejaVuSans.ttf 0.928571
- wqy-microhei.ttc 1.000000
- PionilliumText22L-Medium.ttf 1.000000
import [libs/autoload.lua]: libs/autoload.lua did not return anything
import [ui/FileDialog.lua]: ui/FileDialog.lua did not return anything
Errore di segmentazione (core dump creato)

I tried to clone again from remote and compile, and I tried to rename my ~/.pioneer directory to exclude my settings.

My pioneer version (and OS):

ver git (8e90cc7)
Release(Kernel) Version: 4.13.3-1-ARCH
Kernel Build Timestamp: #1 SMP PREEMPT Thu Sep 21 20:33:16 CEST 2017
Machine Arch: x86_64
NVIDIA nasty proprietary drivers ver. 384.90
KDE Plasma 5.10 as DE

The text was updated successfully, but these errors were encountered:

ColdSpirit0 · 2017-10-02T05:42:36Z

Hmm... works good on my 4.12.13-1-ARCH. Tried to compile exact same version of pioneer and it runs without fails for me.
@Tichy, maybe path to pioneer directory have some no-english characters? If so, try to move it like to /data/pioneer/ and run from here.

If error still exists, please change line 172 at src/LuaUtils.cpp from #undef DEBUG_IMPORT to #define DEBUG_IMPORT, make again and put output here.

Philbywhizz · 2017-10-02T11:16:16Z

I'm having a similar issue with the binary on OSX (compiles ok, blank white window on start and then fails). I've bisected the commit down to commit 5683029

Tichy · 2017-10-02T16:43:20Z

Here's the full log after e enabling DEBUG_IMPORT

https://gist.github.com/Tichy/54d6ee4a816953b1a22522c18f01c056

fluffyfreak · 2017-10-02T16:54:20Z

First thing @Tichy ono-sendai :D

Second thing @ColdSpirit0:

import [libs/autoload.lua]: file libs/autoload.lua loaded
import [libs/autoload.lua]: entering into cache... aborted because imported module returned nil
import [libs/autoload.lua]: generating error messages... got error: libs/autoload.lua did not return anything
import [libs/autoload.lua]: import status: fail

Tichy · 2017-10-02T17:11:15Z

Heh... 😉

jaj22 · 2017-10-02T17:51:05Z

I think the autoload.lua fail is working as intended. Returning nil is used as a signal that the lua file should not be used as an import.

However, there's an assumption in the import code that files will return exactly one value on the lua stack, while pi_lua_dofile will load files which return zero or multiple values. GalacticView.lua returns nothing, so the lua_isnil check is invalid there and so is the pop.

pi_lua_dofile, returns success with any number of return values:

pioneer/src/LuaUtils.cpp

Line 838 in a3aeead

    
           static void pi_lua_dofile(lua_State *l, const FileSystem::FileData &code, int nret)

_import assuming single return value:

pioneer/src/LuaUtils.cpp

Line 504 in a3aeead

// usually happens when file not returns nothing or nil

ColdSpirit0 · 2017-10-02T18:03:04Z

@Tichy thanks!

@fluffyfreak it common behaviour; for import it fail, but for caller it mabe ok. Also autoload.lua should not return anything, it is written in that file. Even if I tried to cache nil, I cant compare returned from cache element, it is not exists or 'nil' (not existing elements I getting like nil value).

@jaj22 thanks, I will check it tomorrow

That debug log from MacOS build (end of log):

import [Lang]: import started, caller: libs/Equipment.lua
import [Lang]: trying to load from cache... not cached
import [Lang]: trying to load a file...Assertion failed: (ty == FileInfo::FT_FILE), function ReadFile, file FileSystemPosix.cpp, line 158.

jaj22 · 2017-10-02T20:42:08Z

The MacOS assert looks like a different issue. I can't find any documentation on stat_fd, which is a bad sign. I think fstat provides that functionality in POSIX.

ColdSpirit0 · 2017-10-03T04:15:14Z

Stack trace from MacOS build:

Thread 0 Crashed:: Dispatch queue: com.apple.main-thread
0   libsystem_kernel.dylib        	0x00007fff8a925f06 __pthread_kill + 10
1   libsystem_pthread.dylib       	0x00007fff941a54ec pthread_kill + 90
2   libsystem_c.dylib             	0x00007fff850786df abort + 129
3   libsystem_c.dylib             	0x00007fff8503fdd8 __assert_rtn + 321
4   pioneer                       	0x000000010990f07e FileSystem::FileSourceFS::ReadFile(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&) + 718 (FileSystemPosix.cpp:158)
5   pioneer                       	0x00000001095e2361 FileSystem::FileSourceUnion::ReadFile(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&) + 65 (.SmartPtr.h:33)
6   pioneer                       	0x0000000109659a44 load_file(lua_State*, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&) + 36 (LuaUtils.cpp:313)
7   pioneer                       	0x0000000109657921 _import(lua_State*, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, bool) + 2705 (LuaUtils.cpp:352)
8   pioneer                       	0x00000001096584cd l_base_import(lua_State*) + 93 (LuaUtils.cpp:538)
9   pioneer                       	0x00000001098f4c86 luaD_precall + 534 (ldo.c:320)
10  pioneer                       	0x0000000109904f0a luaV_execute + 2970 (lvm.c:709)
11  pioneer                       	0x00000001098f5127 luaD_call + 71 (ldo.c:396)
12  pioneer                       	0x00000001098f4656 luaD_rawrunprotected + 86 (ldo.c:134)
13  pioneer                       	0x00000001098f5878 luaD_pcall + 56 (ldo.c:595)
14  pioneer                       	0x00000001098efe0a lua_pcallk + 234 (lapi.c:949)
15  pioneer                       	0x000000010965901f pi_lua_dofile(lua_State*, FileSystem::FileData const&, int) + 223 (LuaUtils.cpp:862)
16  pioneer                       	0x0000000109659a7b load_file(lua_State*, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&) + 91 (.SmartPtr.h:33)
17  pioneer                       	0x000000010965793f _import(lua_State*, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, bool) + 2735 (LuaUtils.cpp:358)
18  pioneer                       	0x00000001096584cd l_base_import(lua_State*) + 93 (LuaUtils.cpp:538)

Also put here my log, without errors: pioneer.log

ColdSpirit0 · 2017-10-03T04:32:44Z

@Tichy your log looks strange for me. It ends on 'success' and no errors after that. Maybe you miss something to copy?

Upd:

That should be after that success, but it does not output...

import [ui/GalacticView.lua]: file ui/GalacticView.lua loaded
import [ui/GalacticView.lua]: entering into cache... aborted because imported module returned nil
import [ui/GalacticView.lua]: generating error messages... got error: ui/GalacticView.lua did not return anything
import [ui/GalacticView.lua]: import status: fail
...

Made a diff of three logs (my, arch, macos) and see that I have only one thread,maybe that why all work on my system? How I can to enable more threads?

jaj22 · 2017-10-03T05:42:06Z

FileDialog.lua also returns nothing (not nil, nothing), so the GalacticView import is already running with a corrupted lua stack. I wouldn't worry about exactly why it's crashing after that. It's inevitable.

Tichy · 2017-10-03T06:56:53Z

@ColdSpirit0 I launched pioneer from both the installed package and the build directory. I can't be sure, but I tend to think that to have something missing I had to delete it myself. In the last test I did nothing but change the #define, make and launch ./pioneer from the src directory.

ColdSpirit0 · 2017-10-03T07:22:13Z

@Tichy ok then, I just asked to be sure)

@jaj22 Im asking for 1 return parameter, I guess pi_lua_dofile puts exact one value, no matter how much values returned .lua file (if no returns, then nil).

pioneer/src/LuaUtils.cpp

Line 316 in a3aeead

pi_lua_dofile(L, *fileData, 1);

Im does not see any logic mistakes in that part of code, but by logs seems it is there.

Bad thing I cant reproduce that errors on my system, just dont know how. Tried to enable any debug flags, but I got lost in 'configure' file.

Upd:

No, debug flags will no help. I edited defines in cpp code and all checks passed.

I will try to install pioneer from AUR...
One more log here: https://aur.archlinux.org/packages/pioneer/
And here: https://aur.archlinux.org/packages/pioneer-git/

Upd 2:

Installed game from https://aur.archlinux.org/packages/pioneer-git/ and all works good.

impaktor · 2017-10-03T08:52:31Z

I'm not following this bug, but was asked for a bt, so here goes (latest master, a3aeead):

font pionillium:
- DejaVuSans.ttf 0.928571
- wqy-microhei.ttc 1.000000
- PionilliumText22L-Medium.ttf 1.000000
import [libs/autoload.lua]: libs/autoload.lua did not return anything
import [ui/FileDialog.lua]: ui/FileDialog.lua did not return anything

Thread 1 "pioneer" received signal SIGSEGV, Segmentation fault.
0x00007ffff73f259a in ?? () from /usr/lib/libSDL2-2.0.so.0
(gdb) bt
#0  0x00007ffff73f259a in ?? () from /usr/lib/libSDL2-2.0.so.0
#1  0x00005555557b2eba in Graphics::TextureBuilder::PrepareSurface (this=this@entry=0x7fffffffd390) at TextureBuilder.cpp:117
#2  0x00005555558b3e3e in Graphics::TextureBuilder::GetDescriptor (this=0x7fffffffd390) at ./../graphics/TextureBuilder.h:48
#3  Graphics::TextureBuilder::CreateTexture (r=0x555555f80770, this=0x7fffffffd390) at ./../graphics/TextureBuilder.h:85
#4  Graphics::TextureBuilder::GetOrCreateTexture (type="ui", r=0x555555f80770, this=0x7fffffffd390) at ./../graphics/TextureBuilder.h:57
#5  UI::Image::Image (this=0x555556e21050, context=<optimized out>, filename=..., sizeControlFlags=12) at Image.cpp:28
#6  0x00005555558e35b4 in UI::Context::Image (sizeControlFlags=12, filename="galaxy_colour.png", this=0x555556ee1260) at ./../ui/Context.h:94
#7  GameUI::GalaxyMap::GalaxyMap (this=0x55555691d730, context=0x555556ee1260) at GalaxyMap.cpp:29
#8  0x00005555558dffb3 in GameUI::LuaGalaxyMap::l_new (l=<optimized out>) at LuaGalaxyMap.cpp:18
#9  0x00005555559730c5 in luaD_precall (L=L@entry=0x555556bbb100, func=<optimized out>, func@entry=0x55555758e220, nresults=nresults@entry=1) at ldo.c:318
#10 0x0000555555981da2 in luaV_execute (L=L@entry=0x555556bbb100) at lvm.c:709
#11 0x000055555597378e in luaD_call (L=0x555556bbb100, func=<optimized out>, nResults=<optimized out>, allowyield=<optimized out>) at ldo.c:395
#12 0x000055555597269f in luaD_rawrunprotected (L=0x555556bbb100, f=0x55555596ab80 <f_call>, ud=0x7fffffffd810) at ldo.c:131
#13 0x0000555555973bd3 in luaD_pcall (L=L@entry=0x555556bbb100, func=func@entry=0x55555596ab80 <f_call>, u=u@entry=0x7fffffffd810, old_top=96, ef=<optimized out>) at ldo.c:595
#14 0x000055555596ff55 in lua_pcallk (L=L@entry=0x555556bbb100, nargs=nargs@entry=0, nresults=nresults@entry=1, errfunc=errfunc@entry=5, ctx=ctx@entry=0, k=k@entry=0x0) at lapi.c:949
#15 0x0000555555692622 in pi_lua_dofile (l=l@entry=0x555556bbb100, code=..., nret=nret@entry=1) at LuaUtils.cpp:862
#16 0x0000555555693a08 in load_file (name="ui/GalacticView.lua", L=0x555556bbb100) at LuaUtils.cpp:316
#17 import_from_file (importInfo=<synthetic pointer>..., L=0x555556bbb100) at LuaUtils.cpp:349
#18 _import (L=L@entry=0x555556bbb100, importName="ui/GalacticView.lua", isFullName=isFullName@entry=true) at LuaUtils.cpp:484
#19 0x000055555569473e in pi_lua_import (isFullName=true, importName="ui/GalacticView.lua", L=0x555556bbb100) at LuaUtils.cpp:545
#20 pi_lua_import_recursive (L=L@entry=0x555556bbb100, basepath="ui") at LuaUtils.cpp:571
#21 0x00005555556b7e24 in LuaInit (this=<optimized out>, this=<optimized out>) at Pi.cpp:321
#22 Pi::Init (options=..., no_gui=<optimized out>) at Pi.cpp:577
#23 0x00005555555c0857 in main (argc=1, argv=<optimized out>) at main.cpp:128
(gdb)

fluffyfreak · 2017-10-03T10:06:39Z

@ColdSpirit0 open your config.ini and make sure that WorkerThreads=0 is set like that.
Do you have a multi-core CPU?
You can also set it to any number you like but by default Pioneer will set it to num_of_cores - 1 so that worker threads won't usually interfere with the main game thread.

Thread:
There appear to be a number of different issues going on here, Arch (Linux), MacOS and a graphics crash.

ColdSpirit0 · 2017-10-03T10:15:10Z

@fluffyfreak thanks. Changing WorkerThreads have no effect to game stability.
Maybe all problems there have one reason)
Tichy and impaktor have same bug.

scottyeager · 2017-10-03T18:35:13Z

I'm also on Arch Linux and Pioneer started crashing after a system update, like so:

import [libs/autoload.lua]: libs/autoload.lua did not return anything
import [ui/FileDialog.lua]: ui/FileDialog.lua did not return anything
fish: 'pioneer' terminated by signal SIGSEGV (Address boundary error)

Downgrading from SDL version 2.0.6 back to 2.0.5 fixed this for me.

laarmen · 2017-10-03T19:11:02Z

I'm guessing pinging @ecraven might be a good idea at this point. Le mar. 3 oct. 2017 à 20:35, scottyeager <notifications@github.com> a écrit :

…

I'm also on Arch Linux and Pioneer started crashing after a system update, like so: import [libs/autoload.lua]: libs/autoload.lua did not return anything import [ui/FileDialog.lua]: ui/FileDialog.lua did not return anything fish: 'pioneer' terminated by signal SIGSEGV (Address boundary error) Downgrading from SDL version 2.0.6 back to 2.0.5 fixed this for me. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#4174 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAkONWflmZzPBLe2D9lvt5CqTMWAYO9Qks5son5jgaJpZM4Pp8-K> .

jaj22 · 2017-10-04T00:01:51Z

It's a deeper problem than UI code. With SDL 2.0.6, SDL_ConvertSurface is crashing on a simple RGB->RGBA conversion of galaxy_colour.png. I did wonder if the weird/broken surface format code in Pioneer was causing an upwards-compatibility problem, but apparently not.

It's likely that the bug isn't in Pioneer code. I'd need to replicate the issue locally with SDL debug libs to be sure.

The MacOS crash is unrelated. LuaUtils attempts to load imports as files before core, and ReadFile in FileSystemPosix.cpp is apparently opening the "data/lang" folder as import("Lang"). Best solution may be for ReadFile to check that it's not about to open a folder. It already has an assert for that, but it needs to handle arbitrary import names without falling over.

fluffyfreak · 2017-10-04T09:39:36Z

Ah, SDL 2.0.6 was just released the other day.

Tichy · 2017-10-04T09:53:05Z

Il 2017-10-04 11:39 Andrew Copland ha scritto:

Ah, SDL 2.0.6 was just released the other day.

I probably have updated pioneer and sdl at the same time.

fluffyfreak · 2017-10-04T11:47:59Z

Yep it's definitely the SDL_ConvertSurface in SDL 2.0.6, I can get it to happen on Windows using VS2015.

EDIT: under debug (x64) I also get a crash using ~~SDL_BlitSurface~~, oops it's actually SDL_SetSurfaceBlendMode loading the mouse cursor.

mooskagh · 2017-10-04T11:50:24Z

Can anyone build/install a SDL with symbol information get a stack trace (which includes line numbers in libsdl)? I'll do that in ~6 hours, but maybe someone will have a chance to do that earlier?

mooskagh · 2017-10-04T17:01:56Z

SDL part of stack trace:

#0  0x00007fd88c3048c6 in SDL_ConvertSurface_REAL (surface=0x55fe047c2840, format=0x55fe01f0a820, flags=0) at /home/crem/tmp/SDL2-2.0.6/src/video/SDL_surface.c:936
#1  0x00007fd88c24251a in SDL_ConvertSurface (a=0x55fe047c2840, b=0x55fe01f0a820, c=0) at /home/crem/tmp/SDL2-2.0.6/src/dynapi/SDL_dynapi_procs.h:492

Context of line 936:

 935     /* Save the original copy flags */
 936     copy_flags = surface->map->info.flags;
 937     copy_color.r = surface->map->info.r;

And surface->map is nullptr.

I'll look further in ~1 hour.

mooskagh · 2017-10-04T17:45:45Z

Seems like a bug in SDL, which was fixed here https://bugzilla.libsdl.org/show_bug.cgi?id=3852 after 2.0.6 release.

UPD: and indeed applying https://hg.libsdl.org/SDL/rev/a09c3f87a12f fixes it.

fluffyfreak · 2017-10-04T18:42:20Z

@mooskagh great work!
I'll have to wait for a point release with the fix before updating pioneer-thirdparty unless I wanna apply that patch and build DLLs/Libs for Windows versions.
Lets hope that they get it out fast too since at the moment I guess distro's are updating too a broken version.

mooskagh · 2017-10-04T18:50:09Z

It's actually @jaj22 who did most of debugging, so credits should go to him. :)

There was a suggestion on #SDL to display a error message if we detect unpatched SDL 2.0.6, so that users pinged their distro maintainers to port the fix.

jaj22 · 2017-10-04T20:33:50Z

@fluffyfreak

EDIT: under debug (x64) I also get a crash using SDL_BlitSurface, oops it's actually SDL_SetSurfaceBlendMode loading the mouse cursor.

I couldn't replicate this, so you may want to check whether it still happens with #4177. x64 crashes in the same place as x86 for me (with SDL 2.0.6).

impaktor · 2017-10-05T06:23:43Z

@mooskagh / crem posted: https://discourse.libsdl.org/t/bug-3852-in-2-0-6/23182

regarding if we should expect a bugfix release.

Tichy · 2017-10-06T17:04:42Z

Confirmed: all fine now 👍

zonkmachine · 2017-10-14T20:26:14Z

@mooskagh / crem posted: https://discourse.libsdl.org/t/bug-3852-in-2-0-6/23182

regarding if we should expect a bugfix release.

SDL 2.0.7 in the works:
https://discourse.libsdl.org/t/sdl-2-0-7-prerelease/23232

fluffyfreak · 2017-10-15T08:57:59Z

@zonkmachine thanks, I'll hold off updating the pioneer-thirdparty repo' until it's dropped then.

zonkmachine · 2017-10-25T01:55:38Z

SDL version 2.0.7 is released:
https://www.libsdl.org/download-2.0.php

jaj22 mentioned this issue Oct 4, 2017

Don't call SDL_FreeSurface until refcount is zero #4177

Merged

ColdSpirit0 mentioned this issue Oct 5, 2017

isFile check for ReadFile (fix for macos crash) #4176

Merged

impaktor closed this as completed in be265b7 Oct 6, 2017

ColdSpirit0 mentioned this issue Oct 20, 2017

Pioneer crashes on Fedora 27 (beta) #4187

Closed

mooskagh mentioned this issue Oct 25, 2017

Assertion failure on startup #4188

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Pioneer crash at start #4174

Pioneer crash at start #4174

Tichy commented Oct 1, 2017

ColdSpirit0 commented Oct 2, 2017 •

edited

Loading

Philbywhizz commented Oct 2, 2017

Tichy commented Oct 2, 2017

fluffyfreak commented Oct 2, 2017

Tichy commented Oct 2, 2017

jaj22 commented Oct 2, 2017

ColdSpirit0 commented Oct 2, 2017 •

edited

Loading

jaj22 commented Oct 2, 2017

ColdSpirit0 commented Oct 3, 2017 •

edited

Loading

ColdSpirit0 commented Oct 3, 2017 •

edited

Loading

jaj22 commented Oct 3, 2017

Tichy commented Oct 3, 2017 •

edited

Loading

ColdSpirit0 commented Oct 3, 2017 •

edited

Loading

impaktor commented Oct 3, 2017 •

edited

Loading

fluffyfreak commented Oct 3, 2017

ColdSpirit0 commented Oct 3, 2017 •

edited

Loading

scottyeager commented Oct 3, 2017

laarmen commented Oct 3, 2017 via email

jaj22 commented Oct 4, 2017

fluffyfreak commented Oct 4, 2017

Tichy commented Oct 4, 2017 via email

fluffyfreak commented Oct 4, 2017 •

edited

Loading

mooskagh commented Oct 4, 2017

mooskagh commented Oct 4, 2017 •

edited

Loading

mooskagh commented Oct 4, 2017 •

edited

Loading

fluffyfreak commented Oct 4, 2017

mooskagh commented Oct 4, 2017

jaj22 commented Oct 4, 2017

impaktor commented Oct 5, 2017

Tichy commented Oct 6, 2017 •

edited

Loading

zonkmachine commented Oct 14, 2017

fluffyfreak commented Oct 15, 2017

zonkmachine commented Oct 25, 2017

Pioneer crash at start #4174

Pioneer crash at start #4174

Comments

Tichy commented Oct 1, 2017

ColdSpirit0 commented Oct 2, 2017 • edited Loading

Philbywhizz commented Oct 2, 2017

Tichy commented Oct 2, 2017

fluffyfreak commented Oct 2, 2017

Tichy commented Oct 2, 2017

jaj22 commented Oct 2, 2017

ColdSpirit0 commented Oct 2, 2017 • edited Loading

jaj22 commented Oct 2, 2017

ColdSpirit0 commented Oct 3, 2017 • edited Loading

ColdSpirit0 commented Oct 3, 2017 • edited Loading

Upd:

jaj22 commented Oct 3, 2017

Tichy commented Oct 3, 2017 • edited Loading

ColdSpirit0 commented Oct 3, 2017 • edited Loading

Upd:

Upd 2:

impaktor commented Oct 3, 2017 • edited Loading

fluffyfreak commented Oct 3, 2017

ColdSpirit0 commented Oct 3, 2017 • edited Loading

scottyeager commented Oct 3, 2017

laarmen commented Oct 3, 2017 via email

jaj22 commented Oct 4, 2017

fluffyfreak commented Oct 4, 2017

Tichy commented Oct 4, 2017 via email

fluffyfreak commented Oct 4, 2017 • edited Loading

mooskagh commented Oct 4, 2017

mooskagh commented Oct 4, 2017 • edited Loading

mooskagh commented Oct 4, 2017 • edited Loading

fluffyfreak commented Oct 4, 2017

mooskagh commented Oct 4, 2017

jaj22 commented Oct 4, 2017

impaktor commented Oct 5, 2017

Tichy commented Oct 6, 2017 • edited Loading

zonkmachine commented Oct 14, 2017

fluffyfreak commented Oct 15, 2017

zonkmachine commented Oct 25, 2017

ColdSpirit0 commented Oct 2, 2017 •

edited

Loading

ColdSpirit0 commented Oct 2, 2017 •

edited

Loading

ColdSpirit0 commented Oct 3, 2017 •

edited

Loading

ColdSpirit0 commented Oct 3, 2017 •

edited

Loading

Tichy commented Oct 3, 2017 •

edited

Loading

ColdSpirit0 commented Oct 3, 2017 •

edited

Loading

impaktor commented Oct 3, 2017 •

edited

Loading

ColdSpirit0 commented Oct 3, 2017 •

edited

Loading

fluffyfreak commented Oct 4, 2017 •

edited

Loading

mooskagh commented Oct 4, 2017 •

edited

Loading

mooskagh commented Oct 4, 2017 •

edited

Loading

Tichy commented Oct 6, 2017 •

edited

Loading