Unicode (utf-8) rework

* Make the errors clearer. (the return error types are not set because that
  reveals bugs in stage1)
* Make decode and length same operation
* Give location of invalid character
* Always report errors in Utf8View

Author: shawnl

lib/std/json.zig

@@ -2099,7 +2099,7 @@ fn unescapeString(output: []u8, input: []const u8) !void {
                 inIndex += 6;
             } else |err| {
                 // it might be a surrogate pair
-                if (err != error.Utf8CannotEncodeSurrogateHalf) {
+                if (err != error.UnicodeSurrogateHalf) {
                     return error.InvalidUnicodeHexSymbol;
                 }
                 // check if a second code unit is present
@@ -2532,15 +2532,16 @@ pub fn stringify(
                             '\r' => try out_stream.writeAll("\\r"),
                             '\t' => try out_stream.writeAll("\\t"),
                             else => {
-                                const ulen = std.unicode.utf8ByteSequenceLength(value[i]) catch unreachable;
                                 // control characters (only things left with 1 byte length) should always be printed as unicode escapes
-                                if (ulen == 1 or options.string.String.escape_unicode) {
-                                    const codepoint = std.unicode.utf8Decode(value[i .. i + ulen]) catch unreachable;
-                                    try outputUnicodeEscape(codepoint, out_stream);
+                                if ((value[i] < 128) or options.string.String.escape_unicode) {
+                                    const c = std.unicode.utf8Decode(value[i..]) catch unreachable;
+                                    try outputUnicodeEscape(c.codepoint, out_stream);
+                                    i += c.utf8len - 1;
                                 } else {
+                                    var ulen = std.unicode.utf8ByteSequenceLength(value[i]) catch unreachable;
                                     try out_stream.writeAll(value[i .. i + ulen]);
+                                    i += ulen - 1;
                                 }
-                                i += ulen - 1;
                             },
                         }
                     }

lib/std/process.zig

@@ -156,7 +156,7 @@ pub const GetEnvVarOwnedError = error{
 pub fn getEnvVarOwned(allocator: *mem.Allocator, key: []const u8) GetEnvVarOwnedError![]u8 {
     if (builtin.os.tag == .windows) {
         const result_w = blk: {
-            const key_w = try std.unicode.utf8ToUtf16LeWithNull(allocator, key);
+            const key_w = std.unicode.utf8ToUtf16LeWithNull(allocator, key) catch return error.InvalidUtf8;
             defer allocator.free(key_w);
 
             break :blk std.os.getenvW(key_w) orelse return error.EnvironmentVariableNotFound;