Generate code that is friendlier to rustc's item_bodies_checking

Having the Some() constructor in each match arm body puts extras stress
on rustc's item_bodies_checking pass. We can work around that by moving
the Some() constructor around the match, and directly returning None
from the default arm, which is the only one that generates a None value.

On my box, this almost cuts the time spent in item_bodies_checking in
half, going from about 8.7s to about 4.6s, and reduces the complete
compile time from about 13s to about 9s, so about a third less.

There are no changes in performance in `cargo bench`.

Note that doing the same for the composition_table() function does not
yield any compile time wins, but would cause a performance regression.

cc #29

Author: dotdash

scripts/unicode.py

@@ -315,14 +315,19 @@ def gen_decomposition_tables(canon_decomp, compat_decomp, out):
     for table, name in tables:
         out.write("#[inline]\n")
         out.write("pub fn %s_fully_decomposed(c: char) -> Option<&'static [char]> {\n" % name)
-        out.write("    match c {\n")
+        # The "Some" constructor is around the match statement here, because
+        # putting it into the individual arms would make the item_bodies
+        # checking of rustc takes almost twice as long, and it's already pretty
+        # slow because of the huge number of match arms and the fact that there
+        # is a borrow inside each arm
+        out.write("    Some(match c {\n")
 
         for char, chars in sorted(table.items()):
             d = ", ".join("'\u{%s}'" % hexify(c) for c in chars)
-            out.write("        '\u{%s}' => Some(&[%s]),\n" % (hexify(char), d))
+            out.write("        '\u{%s}' => &[%s],\n" % (hexify(char), d))
 
-        out.write("        _ => None,\n")
-        out.write("    }\n")
+        out.write("        _ => return None,\n")
+        out.write("    })\n")
         out.write("}\n")
         out.write("\n")