Unicode (and [autoexec]) improvements

[FeralChild64]

Since my previous [autoexec]-related PR was just merged, here is another one :) As the 1st commit is just a file formatting cleanup, and the 2nd one just adds Unicode-related definitions, I suggest to review this PR commit by commit.

---

Unicode related improvement:

• the engine can now also convert strings from DOS code page to UTF8
• the engine can now handle non-normalized Unicode input

[autoexec] section improvements:

• CONFIG -axadd <line> command now converts the line from current DOS code page to UTF-8
• CONFIG -axtype now converts the [autoexec] section to currently set DOS code page

---

Remarks:

• the easiest way to test non-normalized Unicode support is to alter encode.sh or encode.batscript (change nfc to nfkd), and execute it to generate incorrectly normalized translation files; such translations should now work
• the Z:\AUTOEXEC.BAT is not translated from UTF-8 to DOS code page yet (only CONFIG -axtype works with UTF-8); implementing this correctly needs some significant effort - my branch https://github.com/dosbox-staging/dosbox-staging/tree/fc/autoexec-1 mostly does it, but there are still some edge cases which are not resolved yet
• to test the CONFIG -axtype, try the following sequence of command:

KEYB pl 667
CONFIG -axadd ąćęłńóśż
CONFIG -axtype
KEYB pl 668
CONFIG -axtype
KEYB us 437
CONFIG -axtype

To type accented letters above - in keyboard layout pl press Right ALT+letter key combination. Explanation: both 667 and 668 are code pages for Polish language, but mutually incompatible. The 437 does not support Polish, most needed accented characters are absent.

[ThomasEricB]

This is pretty good @FeralChild64 !

[kcgen]

Even though all this Unicode management is a sophisticated task, the code is extremely natural to read with very clear variable and function names, bool names, and extensive self-documenting lambas (along with all the error logging in the labdas to avoid repetition).

This is quite a masterpiece of Unicode handling @FeralChild64!

All of my comments are just minor simplifications or suggestions.

I'm not sure how to type the Polish characters (as I routinely don't use the `Alt+combos), however at least testing w/ the sanitizers and passing in batch files is OK:

If there are any copy-and-paste (or bat files) I can help test with, happy to run them through if you can attach any.

[FeralChild64]

@kcgen

I'm not sure how to type the Polish characters

Right ALT + letter (has to be the right one, left ALT won't work).

[kcgen]

Time to merge; all comments addressed and passing tests. 🚀