[libstd_unicode] Change UNICODE_VERSION to use u32

Use `u32` for version components, as `u64` is just an overkill, and
`u32` is the default type for integers and the default type used for
regular internal numbers.

There's no expectation for Unicode Versions to even reach one thousand
in the next hundered years. This is different from *package versions*,
which may become something auto-generated and exceed human-friendly
range of integer values.

Author: behnam

src/libstd_unicode/tables.rs

@@ -14,7 +14,7 @@
 
 /// The version of [Unicode](http://www.unicode.org/)
 /// that the unicode parts of `CharExt` and `UnicodeStrPrelude` traits are based on.
-pub const UNICODE_VERSION: (u64, u64, u64) = (10, 0, 0);
+pub const UNICODE_VERSION: (u32, u32, u32) = (10, 0, 0);
 
 
 // BoolTrie is a trie for representing a set of Unicode codepoints. It is

src/libstd_unicode/unicode.py

@@ -562,7 +562,7 @@ def emit_norm_module(f, canon, compat, combine, norm_props):
         rf.write("""
 /// The version of [Unicode](http://www.unicode.org/)
 /// that the unicode parts of `CharExt` and `UnicodeStrPrelude` traits are based on.
-pub const UNICODE_VERSION: (u64, u64, u64) = (%s, %s, %s);
+pub const UNICODE_VERSION: (u32, u32, u32) = (%s, %s, %s);
 """ % unicode_version)
         (canon_decomp, compat_decomp, gencats, combines,
                 to_upper, to_lower, to_title) = load_unicode_data("UnicodeData.txt")