OP solved it, closing

.

In general it's not possible to run random tests locally since they are not available until you solve the kata. Ideally tests logs should be more explicit and give you the input your code fails on. In this kata, NASM tests seem identical to C's tests (on Codewars, afaik NASM tests are always written in C), so if you solved them in the latter you may be able to do something useful locally.

My NASM solution passes Sample and Basic tests, but fails with "Signal code: 11" on Random_Tests_1000_strings_of_65635_char

If I understand correctly these tests run various 65635-long strings 1000 times? How could I possibly test something like this by myself?

Hi,

just actually put the fixed tests in the sample tests. No user will try to debug their code with random parameters in the sample tests, expecially because the user doesn't know what your function helper is doing exactly. So actual fixed test are way more useful.

Okay :c
It's not first Shell kata with same problem for me – optimization
Will unlock solutions then, maybe I don't know some crucial tools for such problems
Thank you for kata, I really enjoyed it :з

Sorry, the time out is fixed by CW. Furthermore I can't reduce the tests when some people are timing out...
11 people passed the Shell kata so it's doable in the given time. There is a solution that takes 1275 ms and there are some below 2000 ms. Cheers.

Wow! Because of this even legit but not optimal solutions are invalidated!

(*・‿・)ノ⌒*:･ﾟ✧･ﾟ✧･ﾟ✧･ﾟ✧･ﾟ✧･ﾟ✧･ﾟ✧
            ✧･ﾟ✧magic of coding･ﾟ✧･ﾟ
            ･ﾟ✧･ﾟ✧･ﾟ✧･ﾟ✧･ﾟ✧･ﾟ✧･ﾟ✧

1. Added postmortem log. Looks like this:

Amoeba died from starvation :c
~~~~~~HELP IN DEBUGGING~~~~~~

You've reached maximum amount of iterations: 20
Here is how maze looked like at start:
52 58 64 || :: 
46 || 70 || 94 
40 || 76 82 88 
34 || || || || 
~* 22 16 10 04 
Here is how it looks at the end:
52 58 64 || :: 
46 || 70 || 94 
40 || 76 82 88 
34 || || || || 
~* 22 16 10 04 
You could debug this maze by putting following list in "maze" parameter in test_amoeba_brain():
[[0, 0, 0, 1, 3], [0, 1, 0, 1, 0], [0, 1, 0, 0, 0], [0, 1, 1, 1, 1], [2, 0, 0, 0, 0]]

2. Now user able to provide maze map that looks like this:

maze = [
    [0, 0, 3],
    [0, 1, 1],
    [0, 1, 2]
]
# 0 - empty, 1, wall, 2 -amoeba, 3 - food

More than this - user can copy-paste initial maze structure from postmortem log and try it in sample test.

Added all this in the description of kata.

The fixed tests and random tests are currently very unhelpful in debugging, they might as well not exist. At least these two things are needed:

1. If a random test failed, print out the initial maze state.
2. Change the sample test setup: it should receive a map and use that map as the initial state. Currently it's impossible to even know what map is being tested because it's hidden inside the opaque function.