From d05aec9b66656f8996a04db3be09e91a7b5f3343 Mon Sep 17 00:00:00 2001
From: Kitteh <kitteh@kitteh.pw>
Date: Fri, 4 Jun 2021 13:09:08 +0100
Subject: [PATCH] Use proper utf8 and utf16BE encoding/decoding.

---
 src/qtshit/read/readQVariantT.zig |  3 ++-
 src/qtshit/read/readString.zig    | 33 +++++++++++++------------------
 src/qtshit/utils/unicode.zig      |  2 +-
 src/qtshit/write/writeString.zig  |  7 +++++--
 4 files changed, 22 insertions(+), 23 deletions(-)

diff --git a/src/qtshit/read/readQVariantT.zig b/src/qtshit/read/readQVariantT.zig
index 810b18f..cdfa932 100644
--- a/src/qtshit/read/readQVariantT.zig
+++ b/src/qtshit/read/readQVariantT.zig
@@ -6,6 +6,7 @@ const readShort = @import("./readShort.zig").readShort;
 const readByte = @import("./readByte.zig").readByte;
 const readSignedByte = @import("./readSignedByte.zig").readSignedByte;
 const readQByteArray = @import("./readQByteArray.zig").readQByteArray;
+const ReadStringErrors = @import("./readString.zig").ReadStringErrors;
 const readString = @import("./readString.zig").readString;
 const readQVariantList = @import("./readQVariantList.zig").readQVariantList;
 const readQVariant = @import("./readQVariant.zig").readQVariant;
@@ -17,7 +18,7 @@ const readUserType = @import("./usertypes/readUserType.zig").readUserType;
 const QVariant = @import("../types/QVariant.zig").QVariant;
 const QVariantTypes = @import("../types/QVariantTypes.zig").QVariantTypes;
 
-pub fn readQVariantT(reader: anytype, type_id: u32, allocator: *std.mem.Allocator) (@TypeOf(reader).Error || error{EndOfStream} || error{OutOfMemory} || error{DecodeError})!QVariant {
+pub fn readQVariantT(reader: anytype, type_id: u32, allocator: *std.mem.Allocator) (@TypeOf(reader).Error || error{EndOfStream} || error{OutOfMemory} || error{DecodeError} || ReadStringErrors ) !QVariant {
     switch (type_id) {
         @enumToInt(QVariantTypes.Byte) => {
             var byte = try readByte(reader);
diff --git a/src/qtshit/read/readString.zig b/src/qtshit/read/readString.zig
index c8137e0..83a2c50 100644
--- a/src/qtshit/read/readString.zig
+++ b/src/qtshit/read/readString.zig
@@ -1,30 +1,25 @@
 const std = @import("std");
 const readInt = @import("./readInt.zig").readInt;
-const readShort = @import("./readShort.zig").readShort;
+const range = @import("../utils/RangeIter.zig").range;
+const unicode = @import("../utils/unicode.zig");
 
-pub fn readString(reader: anytype, allocator: *std.mem.Allocator) ![]u8 {
-    var data = std.ArrayList(u8).init(allocator);
-    defer data.deinit();
+pub const ReadStringErrors = (error{ExpectedSecondSurrogateHalf} || error{UnexpectedSecondSurrogateHalf} || error{DanglingSurrogateHalf} || error{OutOfMemory});
 
-    var length = try readInt(reader);
-    var chars = @divTrunc(length, 2);
-    //std.debug.print("read: readString length={d} \n", .{length});
+pub fn readString(reader: anytype, allocator: *std.mem.Allocator) (@TypeOf(reader).Error || error{EndOfStream} || ReadStringErrors)![]u8 {
+    var utf16Data = std.ArrayList(u16).init(allocator);
+    defer utf16Data.deinit();
 
-    var index: usize = 0;
-    while (true) {
-        if (index == chars) break;
+    var num_bytes = @divTrunc(try readInt(reader), 2);
 
-        const byte = try readShort(reader);
-        try data.append(@truncate(u8, byte));
-        index += 1;
+    var iter = range(i32, 0, num_bytes);
+    while (iter.next()) |i| {
+        if (i == num_bytes) break;
+        try utf16Data.append(try reader.readIntLittle(u16));
     }
 
-    var ut8Str = try allocator.alloc(u8, @intCast(usize, chars));
-    for (data.items) |char, i| {
-        ut8Str[i] = char;
-    }
+    var utf8 = try unicode.utf16BEToUtf8(allocator, utf16Data.items);
 
-    //std.debug.print("string: {s}\n", .{ut8Str});
+    //std.debug.print("string: {s}\n", .{utf8});
 
-    return ut8Str;
+    return utf8;
 }
diff --git a/src/qtshit/utils/unicode.zig b/src/qtshit/utils/unicode.zig
index acd11d1..bcaa5a2 100644
--- a/src/qtshit/utils/unicode.zig
+++ b/src/qtshit/utils/unicode.zig
@@ -14,7 +14,7 @@ pub fn utf8ToUtf16BE(allocator: *std.mem.Allocator, utf8: []const u8) ![]u16 {
     var utf16LE = try std.unicode.utf8ToUtf16LeWithNull(allocator, utf8);
     // Little Endian to Big Endian
     var utf16BE = byteSwapArray(u16, &utf16LE);
-    return utf16BE;
+    return utf16BE[0..utf16BE.len];
 }
 
 pub fn utf16BEToUtf8(allocator: *std.mem.Allocator, utf16: []u16) ![]u8 {
diff --git a/src/qtshit/write/writeString.zig b/src/qtshit/write/writeString.zig
index 8a0fe68..f05139a 100644
--- a/src/qtshit/write/writeString.zig
+++ b/src/qtshit/write/writeString.zig
@@ -2,10 +2,13 @@ const std = @import("std");
 const writeInt = @import("./writeInt.zig").writeInt;
 const unicode = @import("../utils/unicode.zig");
 
+pub const WriteStringErrors = (error{InvalidUtf8} || error{OutOfMemory});
 
-pub fn writeString(writer: anytype, allocator: *std.mem.Allocator, str: []const u8) (error{InvalidUtf8} || error{OutOfMemory})!void {
+pub fn writeString(writer: anytype, allocator: *std.mem.Allocator, str: []const u8) WriteStringErrors!void {
     var str_utf16BE = try unicode.utf8ToUtf16BE(allocator, str);
     defer allocator.free(str_utf16BE);
-    try writeInt(writer, @intCast(i32, str_utf16BE.len));
+
+    // len * 2 because length of u8 not length of u16.
+    try writeInt(writer, @intCast(i32, str_utf16BE.len * 2));
     try writer.writeAll(std.mem.sliceAsBytes(str_utf16BE));
 }
\ No newline at end of file