JOURNAL.md

Day 1

It's a mistake coding in C. Nevertheless, I managed to scrape through the whole thing. Admittedly, I overcomplicate, overengineer, and overefficientize everything. So it's mostly my fault.

In theory, this whole day's AoC could have been done using vim or sed + bc

1-1

The theory behind 1-1's code is simple: in that it's not.

1. Read the file line by line (max 64 char per line).
2. Turn all encountered digits to their int value ('0' -> 0, done by subtracting '0' or ascii val 30).
3. Bit shift to the left all encountered digits by 4*k where k is the amount of digits encountered starting from 0 (the logic being that the value 9 occupies only 4 bits).
4. Modulo z by 1 << 4 to get the first digit (and * 10 to make it the tenths).
5. Divide z by 1 << (4 * (k - 1)) to get the last digit.
6. Add the previous two together, add all the lines, and you get the total result.

1-2

1-2's code is blatant madness. Here's how the "logic" (if any) goes:

1. Read the file line by line (max 64 char per line).
2. Iterate through the line char by char.
3. If a digit is encountered:
   1. Turn it into a number (by subtracting '0').
   2. Set xc or yc to 1 (essentially locking it).
4. If any character in "otfsen" is found, then numericmagify:
   1. Check len * 5 and prevent it from exceeding sizeof(uint_fast32_t) * 8, if exceeding, use sizeof(uint_fast32_t) * 3 / 2 instead.
   2. Then, for every char in str, subtract by '`' so that 'a' = 1, then bit shift left by 5*i (the logic being 'z' - '`' is 5 bits long).
   3. Now check if the encountered chars really constitutes a number by checking the "magic"/"hash" against the number-word you want.
   4. If yes, set the number to x or y and set xc or yc to 1 (essentially locking it).

Note: the encounter is done from the front and behind of the buf even if it's 1 loop.
Note2: yes, numericmagify is a shabby, knockoff hash function.

Day 2