Andrew Reece
51 posts / 3 projects
Maker of things
Immediate crash on opening (v0.7.2 Win7 x64)
Edited by Andrew Reece on Reason: Initial post
Hi,

I just bought this on itch.io. v0.7.2 (the latest at the time of posting) crashes immediately on opening with the
 1 codeclap has stopped working 
message.

v0.6.4 opens without a problem.

Let me know if I can provide any other info.

Cheers,
Andrew
148 posts / 1 project
Immediate crash on opening (v0.7.2 Win7 x64)
That's not a known issues so I'd love to get a few hints on where to look for this.

A small list of things that come to my mind right away:
• I guess you probably have tried executing from a clean directory without overwriting any files?
• Are you using some command line parameters or are you simply double clicking the .exe in the explorer?
• If you have visual studio installed windows should give you the option to debug the stopped program. There you should be able to see the function and offset into the function where the fault occurred, that could be one of the most helpful hints.

Thanks
Andrew Reece
51 posts / 3 projects
Maker of things
Immediate crash on opening (v0.7.2 Win7 x64)
I unzipped the file and double-clicked the exe inside.

There's no PDB in the folder, so VS hasn't loaded symbols.

I'll include the details I can give you below, but I'm not sure how helpful they are... If you want to give me the PDB, I'm happy to be more specific.

Disassembly
  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26  000000013FE74B60 mov qword ptr [rsp+8],rcx 000000013FE74B65 sub rsp,38h 000000013FE74B69 mov rdx,qword ptr [rsp+40h] 000000013FE74B6E mov cx,1 000000013FE74B72 call 000000013FE74A50 000000013FE74B77 mov qword ptr [rsp+20h],rax 000000013FE74B7C cmp qword ptr [rsp+20h],0 000000013FE74B82 je 000000013FE74BB7 000000013FE74B84 mov rax,qword ptr [rsp+20h] 000000013FE74B89 mov rax,qword ptr [rax] 000000013FE74B8C mov qword ptr [rsp+28h],rax 000000013FE74B91 call 000000013FE73D50 000000013FE74B96 mov rax,qword ptr [rax+30h] 000000013FE74B9A mov rcx,qword ptr [rsp+28h] > 000000013FE74B9F cmp qword ptr [rcx+78h],rax 000000013FE74BA3 jne 000000013FE74BB1 000000013FE74BA5 mov rax,qword ptr [rsp+20h] 000000013FE74BAA mov rax,qword ptr [rax] 000000013FE74BAD jmp 000000013FE74BB9 000000013FE74BAF jmp 000000013FE74BB5 000000013FE74BB1 xor eax,eax 000000013FE74BB3 jmp 000000013FE74BB9 000000013FE74BB5 jmp 000000013FE74BB9 000000013FE74BB7 xor eax,eax 000000013FE74BB9 add rsp,38h 000000013FE74BBD ret 

Call stack
 1 2 3 4 5 6 7 8 > codeclap.exe!000000013fe74b9f() Unknown codeclap.exe!000000013fe75176() Unknown codeclap.exe!000000013fe764d6() Unknown codeclap.exe!000000013fe49208() Unknown codeclap.exe!000000013fed4548() Unknown codeclap.exe!000000013fee0358() Unknown kernel32.dll!00000000775d59cd() Unknown ntdll.dll!000000007783385d() Unknown 

Exception error
 1 Unhandled exception at 0x000000013FE74B9F in codeclap.exe: 0xC0000005: Access violation reading location 0x0000000000000078. 
148 posts / 1 project
Immediate crash on opening (v0.7.2 Win7 x64)
Edited by spx on
ah seems I wrongly assumed that most function addresses are also still present in the Windows build (even without the full debug information).
I've uploaded the pdbs to itch.io and they should match the release build (I hope, after switching to msvc I never actually made sure that this is true). Just extract/copy them into the same directory and you should get a more explanatory trace.
Andrew Reece
51 posts / 3 projects
Maker of things
Immediate crash on opening (v0.7.2 Win7 x64)
Thanks. The crash is in this file:

 1 z:\impromptu\ws2\codeclap\iui.c 

and looks like it's on line 1879.

Call stack:
 1 2 3 4 5 6 7 > codeclap.exe!iuiCurrentPanel(iui_command * IgnoreCommand) Line 1879 C codeclap.exe!iuiGenericWidgetBegin(iui_type Type, iui_rect Margin, iui_rect Padding) Line 2042 C codeclap.exe!iuiGenericWidget(iui_type DrawType, iui_state DrawState, iui_rect Margin, iui_rect Padding, float ForceWidth, const char * Label, iui_align Align) Line 2361 C codeclap.exe!iuiRadioButton(const char * Label, iui_type Type, int * Option) Line 3798 C codeclap.exe!DebuggerFrontendProcess(debugger_input * Input) Line 13942 C codeclap.exe!CodeclapFrontendThread(void * lpParam) Line 975 C [External Code] 
148 posts / 1 project
Immediate crash on opening (v0.7.2 Win7 x64)
Thanks, the debug information matches with the information you have provided before.
I may have a clue whats going on here and will have to do some checking on different machines tomorrow.

One additional thing that may be helpful, what CPU are you using?
Mārtiņš Možeiko
2405 posts / 2 projects
Immediate crash on opening (v0.7.2 Win7 x64)
spx
ah seems I wrongly assumed that most function addresses are also still present in the Windows build (even without the full debug information).

For that you typically build with "/PDBSTRIPED:second_file.pdb" argument. This will contain only public symbols + FPO records for stack walking, which you can distribute with your binaries (it is also smaller). You can use this argument in additional to normal "/PDB:file.pdb" argument - it basically creates two pdb files, one for you, one for distribution.
Andrew Reece
51 posts / 3 projects
Maker of things
Immediate crash on opening (v0.7.2 Win7 x64)
spx
Thanks, the debug information matches with the information you have provided before.
I may have a clue whats going on here and will have to do some checking on different machines tomorrow.

One additional thing that may be helpful, what CPU are you using?

No worries, glad to hear it.

It's an Intel i5-3570K, which I think is Ivy Bridge.
148 posts / 1 project
Immediate crash on opening (v0.7.2 Win7 x64)
Edited by spx on
azmr
No worries, glad to hear it.
It's an Intel i5-3570K, which I think is Ivy Bridge.

Thank you for the quick report. During transit to msvc I replaced usage of bsr with lzcnt without putting too much thought into it. On pre-haswell processors this executes (despite not being supported) but produces wrong results leading to this behavior.
148 posts / 1 project
Immediate crash on opening (v0.7.2 Win7 x64)
Edited by spx on
Unfortunately I will have to wait a few weeks to test on an ivy bridge CPU myself but 0.7.5 should work for you again.