Overlays: Fix wrapped text in native UI

[Megamouse]

• The first commit aligns wrapped lines properly instead of only aligning the whole wrapped block as seen in master.
• The second commit fixes the missing characters that sometimes occur at the beginning of the wrapped lines.
  I had to rewrite render_text_ex because I did not really understand the existing code.
  It feels like I simplified the function in the process.

Old:
Notice the missing [ of "[Cross-Controller]" at the beginning of the third line.

New:

[Darkhost1999]

I'm able to write my own messages using Bethesda games. Looks good to me even with multiple spaces before a word. I hadn't tested earlier so IDK if it was broke before

Test Results

PR

Master

[kd-11]

I'm able to write my own messages using Bethesda games. Looks good to me even with multiple spaces before a word. I hadn't tested earlier so IDK if it was broke before
Test Results

That's not what I was talking about. You need multiple spaces at the end of a wrapped line to see the difference.

[Megamouse]

Btw, the original bug is really ugly. You access characters of the original text array with indices that don't match because of newlines etc. So I think my approach is much less convoluted and tracking the indices is even less performant imo.

I don't quite understand what you mean with the whitespace stuff.

Used string:

"           Start [%0] on the PS Vita system.             \nIf you have not installed [%0], go to [Remote Play] on the            PS            Vita            system            and            start            [Cross-Controller] from the LiveArea™ screen.                        "

This is your patched code ( '[' now isn't truncated in the original dialog ):

Mine:

It looks pretty much identical if you ask me.
The only thing both algorithms don't do is truncate the end of the lines.

[Megamouse]

Fixed the alignment with leading and trailing whitespace

[kd-11]

I don't quite understand what you mean with the whitespace stuff.

I meant leading multiple whitespace. You needed to keep ignoring whitespace after a word break until a non-whitespace character was encountered. Seems fine now.