Allow surrogates in string literals.

docs/language/dartLangSpec.tex

Index: docs/language/dartLangSpec.tex
diff --git a/docs/language/dartLangSpec.tex b/docs/language/dartLangSpec.tex
index c1d9e6b06ac936769a15bcd52f42dcfcd577a356..f4904b2ba7bb08009b88e1693626d69ef8701210 100644
--- a/docs/language/dartLangSpec.tex
+++ b/docs/language/dartLangSpec.tex
@@ -2790,7 +2790,7 @@ Strings support escape sequences for special characters. The escapes are:
 
 $\backslash$u\{$HEX\_DIGIT_1$ $HEX\_DIGIT_2$\}.
 \item $\backslash$u $HEX\_DIGIT_1$ $HEX\_DIGIT_2$ $HEX\_DIGIT_3$ $HEX\_DIGIT_4$, equivalent to $\backslash$u\{$HEX\_DIGIT_1$ $HEX\_DIGIT_2$ $HEX\_DIGIT_3$ $HEX\_DIGIT_4$\}.
-\item $\backslash$u\{$HEX\_DIGIT\_SEQUENCE$\} is the unicode scalar value represented by the $HEX\_DIGIT\_SEQUENCE$. It is a compile-time error if the value of the $HEX\_DIGIT\_SEQUENCE$ is not a valid unicode scalar value.
+\item $\backslash$u\{$HEX\_DIGIT\_SEQUENCE$\} is the Unicode code point represented by the $HEX\_DIGIT\_SEQUENCE$.

[floitsch, 2016/09/15 15:09:19]

It might only be the unicode code unit, since we now allow surrogate pairs.

[Lasse Reichstein Nielsen, 2016/09/15 15:50:37]

The term "code unit" presumes an encoding form (e.g., UTF-16 where code units
are 16-bit unsigned integers). Here we are actually talking about the code point
itself. The values in the surrogate range are code points but not Unicode scalar
values (that's basically the definition of "Unicode scalar value": Code points
except high and low surrogates), hence the change.

I should put back the "compile-time error" part, though, since six hex digits
can specify a value which is not a valid code unit - anything above 0x10FFFF.

 \item \$ indicating the beginning of an interpolated expression.
 \item Otherwise, $\backslash k$ indicates the character $k$ for any $k$ not in $\{n, r, f, b, t, v, x, u\}$.
  \end{itemize}