TDPL, page 37, says:

"Hex strings are useful for defining raw data; the compiler makes no attempt whatsoever to interpret the contents as Unicode characters or anything else but hexadecimal digits."

But:

static immutable (char[32])  s0 = x"0000";    // compiles
static string                s1 = x"cccc";    // compiles
static immutable (char[])    s2 = x"cccc";    // compiles
static immutable (char[32])  s3 = x"cccc";    // Error: Invalid trailing code unit
static immutable (char[32])  s4 = x"cc";      // Error: Truncated UTF-8 sequence
static immutable (ubyte[])   s5 = x"cccc";    // Error: cannot implicitly convert ...

static immutable (ubyte[32]) s6 = x"cccc";    // Error: Invalid trailing code unit
                                              // Error: cannot implicitly convert ...

"Invalid trailing code unit" and "Truncated UTF-8 sequence" certainly look like Unicode interpretation errors to me.

(Of course, what I would really like to do is initialize a ubyte array with a hex string, preferably without a cast.)

I cannot find any relevant documentation on either of these pages.
http://dlang.org/lex.html
http://dlang.org/arrays.html#strings

Any advice would be appreciated.  Thanks!