Decide if we should use std::string vs Util::UString vs Util::Buffer

[sppmacd]

std::string is not Unicode-compatible. It is often used for storing binary data or random text (taken from files). Convert all these usages to Util::UString or Util::Buffer. For code-internal strings, use std::string_view or char[S].

[sppmacd]

actually, std::string can be used when we won't display this text anywhere (or only in the console)

Also Util::UString has its own problems, notably it is encoded in UTF-32 which is a mostly a waste of memory, also codepoints are not the only way to define "characters" of a text (we should support graphemes for text widgets).

I think we should just use UTF-8 in most cases, and UTF-32 only if we need fast random access (for TextEditor maybe)

So, for example, we would have

• Util::String for most text (UTF-8 internally, but this is abstracted away, we can iterate over codepoints, graphemes, etc.),
• Util::U32String (UTF-32 internally)
  • why don't just use std::vector<u32> in these rare cases)
• Util::Buffer for binary data

Or other design:

• Use std::string / Util::Buffer / ??? everywhere
• Remember to use some "iterator" classes UTF8Iterator, UTF32Iterator etc based on the underlying encoding, they would allow iterating over codepoints or graphemes or whatever