If only one could look into the future like this to understand if tommorow's prices would rise... IMHO the real final value should be the current value of the shares (i.e. last price) if you're still holding them.

fixed

Tests cannot run in C. (edit: Attempt ran without these errors and failed as expected but Test delivers these errors)

fixture.c:41:33: error: too few arguments to function call, expected 3, have 2
    assert_approx_eq(actual, 1.0);
    ~~~~~~~~~~~~~~~~            ^
fixture.c:17:1: note: 'assert_approx_eq' declared here
static void assert_approx_eq(const double prices[], size_t n_prices, double expected)
^
fixture.c:47:33: error: too few arguments to function call, expected 3, have 2
    assert_approx_eq(actual, 1.5);
    ~~~~~~~~~~~~~~~~            ^
fixture.c:17:1: note: 'assert_approx_eq' declared here
static void assert_approx_eq(const double prices[], size_t n_prices, double expected)
^
fixture.c:53:33: error: too few arguments to function call, expected 3, have 2
    assert_approx_eq(actual, 1.5);
    ~~~~~~~~~~~~~~~~            ^
fixture.c:17:1: note: 'assert_approx_eq' declared here
static void assert_approx_eq(const double prices[], size_t n_prices, double expected)
^
3 errors generated.

Despitee terrible efficiency, funny enough its still O(n). :)
O(25n)==O(n)

done

My rust translation shows both. You get the u8 bytes and the same data decoded as a Latin1 string.

(I meant to mark this suggestion as resolved, hereby).

I've now published my translation, see the new suggestion I made.

I've created a new translation; the tests in the translation were partially based on yours but I extended them (adding the advanced and random tests, as well as several corrections made to the general kata test suites). I also re-used your starting template.

My translation features an optimising BF interpreter that matches and exceeds the optimisations made in the other kata translations here.

I might pick this up at some point. I already started on a Rust version, your version could be helpful in accellerating the process.

I've fixed this across all translations, by escaping the quote:

end
msg " That\'s end"

Got the answer, kept for posterity to clarify the rule on defender bonus calculation:
"M attack. R Power 4. R wins
Go to the trench
P has power 3 of the L team. It gets the trench defense bonus 6 so it wins. 2 remains after the successful defense.
2 points is less than the original 3 so it keeps 2 power
It overpowers first s and losses 1 power point keeping 1.
Overcomes the second s with a draw and retreats. The next s wins its position"

Can someone explain why with this input:

m     
-|pzzz

the expected output is R|LLLR (not R|LLLL)?

Haskell fork that updates the random test generation to avoid generating numbers with exponents.

Sorry for offtop, but what software you used to generate this image? I was looking for something like that for a while