Pretty-print all String-like formats as strings, rather than sequences of characters

[archaephyrryx]

Currently, the ASCII string types we want to display as ascii are:

• Any compound Format whose leaf-values are all base.ascii-char or base.ascii-char.* (glob, not regex)
  • (With the following exceptions: ???)
• "base.asciiz-string" :=
  record([("string", repeat(not_byte(0x00))), ("null", is_byte(0x00))])
• "tar.ascii-string" :=
  record([("string", repeat(not_byte(0x00))), ("padding", repeat1(is_byte(0x00)))])
• "tar.ascii-string.opt0" :=
  record([("string", repeat(not_byte(0x00))), ("padding", repeat(is_byte(0x00)))])
• "tar.ascii-string.opt0.nonempty" :=
  record([("string", repeat1(not_byte(0x00))), ("padding", repeat(is_byte(0x00)))])

Possible classifications

• By direct properties of leaf elements
• By name of interior elements (e.g. Seq, Repeat, Tuple, or Record whose non-implicit elements are all base.ascii-char(.*)?)
• By name pattern-match (e.g. /ascii/ && /string/ or /ascii-string/ || /asciiz-string/)
• By structural pattern match (e.g. simply-nested or top-level record with a field named "string" satisfying certain ascii-like properties)

Possible workarounds:
- Move format-specific string types to Base and enforce similar prefixing logic to base.ascii-char and derivatives. This may require renaming base.asciiz-string to base.ascii-string.cstr or similar

EDIT:

This should extend to Unicode strings as well, as utf8.string = repeat utf8.char is currently handled in sequences rather than string-form. This may be good enough for now, but if text formats are ever unified (i.e. ascii as a subset of UTF-8), this may be more relevant than it is now.

[mikeday]

I'd like to see how far we can get by leveraging ascii-char and building up from there, the only complicated part is that some strings require a nul terminator, some have an optional nul terminator, and some are a fixed width and may include arbitrary nul characters. Perhaps ascii-char, ascii-nul, and ascii-non-nul would be sufficient?

[archaephyrryx]

Now that we have tentative support of UTF-8 based on #99, we might consider broadening the criteria for 'string-treatment' a bit.