Segmentation Fault Core dumped

[marble pulsar]

Been hard stucking on running the follow function as binary in Linux. It can run by mojo run .mojo just fine but it just can’t be executed as a binary … and the only error I got is in the title.

    var ecwords = List[Int](0)*64
    var values = List[Int]()

    for i in range(222):
        var temp: Int = (padded[i] + ecwords[0]) % 929
        for j in range(63, -1, -1):
            var val: Int = temp * EC_LV5[j] % 929
            var factor: Int = ecwords[64-j] if j > 0 else 0
            ecwords[63-j] = (factor + 929 - val) % 929

    return ecwords```
EC_LV5 is 64 bytes of error correction codewords ECL_LV5 = List[Int](….)

[spare token]

Do you have any way of putting this in gdb at the moment? We can try to tease apart where things might be falling over

[silent parrot]

what is the intended value of ecwords?

[marble pulsar]

output value is correct . A list of 64 elements values from 0-929

[spare token]

are you able to run any of this in GDB or some other debugger?

[marble pulsar]

I’ve only tested in Ubuntu terminal and vscode

[snow marlin]

Could you share the full script?

[marble pulsar]

I can run the fn without any issue I couldn’t run it in binary execution form

[spare token]

Yeah, can in the ubuntu VM/wsl can you install gdb?

[marble pulsar]

Like ./test

[marble pulsar]

I’ll try right now

Not VM not WSL Ubuntu OS

[marble pulsar]

I’ll provide result in few mins

  <https://debuginfod.ubuntu.com>
Enable debuginfod for this session? (y or [n]) y
Debuginfod has been enabled.
To make this setting permanent, add 'set debuginfod enabled on' to .gdbinit.
Downloading separate debug info for system-supplied DSO at 0x7ffff7fc3000
Downloading separate debug info for /lib/x86_64-linux-gnu/libstdc++.so.6        
Downloading separate debug info for /lib/x86_64-linux-gnu/libgcc_s.so.1         
[Thread debugging using libthread_db enabled]                                   
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
[New Thread 0x7ffff6e006c0 (LWP 11477)]
[New Thread 0x7ffff64006c0 (LWP 11478)]
[New Thread 0x7ffff5a006c0 (LWP 11479)]
[New Thread 0x7ffff50006c0 (LWP 11480)]
[New Thread 0x7ffff46006c0 (LWP 11481)]

Thread 1 "errc" received signal SIGSEGV, Segmentation fault.
0x000055555555bf6c in main ()

I somehow fixed it by putting var ECL_V5 from outside of the fn into inside of the fn…. Still don’t know what’s the actual reason behind

[spare token]

can you maybe add -g or something to the compiler flags?

need debug symbols to diagnose

[marble pulsar]

I discovered that if a variable is outside of the fn it should be defined as alias not var in order to make the binary work. If its defined as var the script can still run with mojo run

[native pine]

Congrats @marble pulsar, you just advanced to level 1!

[marble pulsar]

Thank you for the gdb knowledge which I didn’t know 🙂

[marble pulsar]

The issue was about defining global variables/constants as var vs. as alias . In my case it needs to be defined as alias for the binary to work

As a dude from python took me a while to figure out. And still reading the variables scope thing

[spare token]

interesting! weird that caused a segfault though? I forget if global vars are supported yet or not

[foggy lynx]

they're not currently supported and the segfault on attempting to use one is a known bug. There probably should be a warning from the lsp and a mention in the 'sharp edges' page