2020-2: Character Building

UTF-8 encodes Unicode characters using 1-4 integers for each character.

Uiua includes a system function, utf₈, that can convert characters into integers and integers into characters.

The expression utf₈ "UTF-8" converts between characters and UTF-8 and °utf₈ does the reverse.

Consider the following:

utf₈ "D¥⍺⌊○9"

[68 194 165 226 141 186 226 140 138 226 151 139 57]

°utf₈ [68 194 165 226 141 186 226 140 138 226 151 139 57]

"D¥⍺⌊○9"

How many integers does each character use?

≡(□utf₈) graphemes "D¥⍺⌊○9"

{[68] [194 165] [226 141 186] [226 140 138] [226 151 139] [57]}

The rule is that an integer in the range 128 to 191 (inclusive) continues the character of the previous integer (which may itself be a continuation).

With that in mind, Write code that, given an argument which is a simple integer vector representing valid UTF-8 text, encloses each sequence of integers that represent a single character, like the result of 'UTF-8'∘⎕UCS¨'UTF-8'∘⎕UCS but does not use any system functions (names beginning with ⎕).