fixed

fixed

glad i found this answer (i did not think about that). thank you, you are great!

approved

description now states:

If an index is beyond the string, it should be ignored.

fixed

This was an excellent little exercise in sorting logic and value mapping.
It really pushed me to think clearly about how to rank character significance and deal with ties deterministically.
Thanks to the kata creator for the challenge!

Python fork with updated tests.

JavaScript fork (kata author inactive)

• Node 18 + chai assertions
• the random tests are no longer susceptible to input mutation (fixing this issue)
• clearly spelling out in the description that out-of-bounds indices should be ignored (fixing this issue)
• the assertion messages now include the input values for better user experience

this must have been fixed at some point, or was a user error. the test case is now:

Test.assertDeepEquals(capitalize("abcdef",[1,2,5]),'aBCdeF');

The description indicates that out-of-bounds indices should be ignored. Your out-of-bounds if guard is misguided:

if (str[i])

This is checking whether str[i] is a nul byte. this will access memory that you should not be accessing, located in addresses beyond str. you have to check that i is within the length of str instead. the strange assertion messages you are seeing are due to you mutating memory that you are not supposed to.

by the way, contrary to what the comment in your code implies, strdup() is available on Codewars. you just have to declare it somewhere with char *stdup(const char *);. this is because strdup() requires C23, and codewars uses C18; but strdup() was available on POSIX and Windows long before C23.

Haskell fork that fixes the random tests so that failed tests show the input strings along with some other minor improvements to test generation.