Merge pull request #311 from lightpanda-io/msg_size_encode

Msg size encode

src/handler.zig

@@ -19,6 +19,7 @@
 const std = @import("std");
 
 const ws = @import("websocket");
+const Msg = @import("msg.zig").Msg;
 
 const log = std.log.scoped(.handler);
 
@@ -45,7 +46,7 @@ pub const Stream = struct {
     }
 
     fn closeCDP(self: *const Stream) void {
-        const close_msg: []const u8 = "5:close";
+        const close_msg: []const u8 = .{ 5, 0 } ++ "close";
         self.recv(close_msg) catch |err| {
             log.err("stream close error: {any}", .{err});
         };
@@ -82,8 +83,10 @@ pub const Handler = struct {
         self.stream.closeCDP();
     }
 
-    pub fn clientMessage(self: *Handler, alloc: std.mem.Allocator, data: []const u8) !void {
-        const msg = try std.fmt.allocPrint(alloc, "{d}:{s}", .{ data.len, data });
-        try self.stream.recv(msg);
+    pub fn clientMessage(self: *Handler, data: []const u8) !void {
+        var header: [2]u8 = undefined;
+        Msg.setSize(data.len, &header);
+        try self.stream.recv(&header);
+        try self.stream.recv(data);
     }
 };

src/main.zig

@@ -25,6 +25,7 @@ const websocket = @import("websocket");
 const Browser = @import("browser/browser.zig").Browser;
 const server = @import("server.zig");
 const handler = @import("handler.zig");
+const MaxSize = @import("msg.zig").MaxSize;
 
 const parser = @import("netsurf");
 const apiweb = @import("apiweb.zig");
@@ -274,6 +275,8 @@ pub fn main() !void {
             var ws = try websocket.Server(handler.Handler).init(alloc, .{
                 .port = opts.port,
                 .address = opts.host,
+                .max_message_size = MaxSize + 14, // overhead websocket
+                .max_conn = 1,
                 .handshake = .{
                     .timeout = 3,
                     .max_size = 1024,

src/msg.zig

@@ -18,44 +18,47 @@
 
 const std = @import("std");
 
-/// MsgBuffer returns messages from a raw text read stream,
-/// according to the following format `<msg_size>:<msg>`.
+pub const MsgSize = 16 * 1204; // 16KB
+pub const HeaderSize = 2;
+pub const MaxSize = HeaderSize + MsgSize;
+
+pub const Msg = struct {
+    pub fn getSize(data: []const u8) usize {
+        return std.mem.readInt(u16, data[0..HeaderSize], .little);
+    }
+
+    pub fn setSize(len: usize, header: *[2]u8) void {
+        std.mem.writeInt(u16, header, @intCast(len), .little);
+    }
+};
+
+/// Buffer returns messages from a raw text read stream,
+/// with the message size being encoded on the 2 first bytes (little endian)
 /// It handles both:
 /// - combined messages in one read
 /// - single message in several reads (multipart)
-/// It's safe (and a good practice) to reuse the same MsgBuffer
+/// It's safe (and a good practice) to reuse the same Buffer
 /// on several reads of the same stream.
-pub const MsgBuffer = struct {
-    size: usize = 0,
+pub const Buffer = struct {
     buf: []u8,
+    size: usize = 0,
     pos: usize = 0,
 
-    const MaxSize = 1024 * 1024; // 1MB
-
-    pub fn init(alloc: std.mem.Allocator, size: usize) std.mem.Allocator.Error!MsgBuffer {
-        const buf = try alloc.alloc(u8, size);
-        return .{ .buf = buf };
-    }
-
-    pub fn deinit(self: MsgBuffer, alloc: std.mem.Allocator) void {
-        alloc.free(self.buf);
-    }
-
-    fn isFinished(self: *MsgBuffer) bool {
+    fn isFinished(self: *const Buffer) bool {
         return self.pos >= self.size;
     }
 
-    fn isEmpty(self: MsgBuffer) bool {
+    fn isEmpty(self: *const Buffer) bool {
         return self.size == 0 and self.pos == 0;
     }
 
-    fn reset(self: *MsgBuffer) void {
+    fn reset(self: *Buffer) void {
         self.size = 0;
         self.pos = 0;
     }
 
     // read input
-    pub fn read(self: *MsgBuffer, alloc: std.mem.Allocator, input: []const u8) !struct {
+    pub fn read(self: *Buffer, input: []const u8) !struct {
         msg: []const u8,
         left: []const u8,
     } {
@@ -64,11 +67,9 @@ pub const MsgBuffer = struct {
         // msg size
         var msg_size: usize = undefined;
         if (self.isEmpty()) {
-            // parse msg size metadata
-            const size_pos = std.mem.indexOfScalar(u8, _input, ':') orelse return error.InputWithoutSize;
-            const size_str = _input[0..size_pos];
-            msg_size = try std.fmt.parseInt(u32, size_str, 10);
-            _input = _input[size_pos + 1 ..];
+            // decode msg size header
+            msg_size = Msg.getSize(_input);
+            _input = _input[HeaderSize..];
         } else {
             msg_size = self.size;
         }
@@ -77,7 +78,7 @@ pub const MsgBuffer = struct {
         const is_multipart = !self.isEmpty() or _input.len < msg_size;
         if (is_multipart) {
 
-            // set msg size on empty MsgBuffer
+            // set msg size on empty Buffer
             if (self.isEmpty()) {
                 self.size = msg_size;
             }
@@ -90,18 +91,7 @@ pub const MsgBuffer = struct {
                 return error.MsgTooBig;
             }
 
-            // check if the current input can fit in MsgBuffer
-            if (new_pos > self.buf.len) {
-                // we want to realloc at least:
-                // - a size big enough to fit the entire input (ie. new_pos)
-                // - a size big enough (ie. current msg size + starting buffer size)
-                // to avoid multiple reallocation
-                const new_size = @max(self.buf.len + self.size, new_pos);
-                // resize the MsgBuffer to fit
-                self.buf = try alloc.realloc(self.buf, new_size);
-            }
-
-            // copy the current input into MsgBuffer
+            // copy the current input into Buffer
             // NOTE: we could use @memcpy but it's not Thread-safe (alias problem)
             // see https://www.openmymind.net/Zigs-memcpy-copyForwards-and-copyBackwards/
             // Intead we just use std.mem.copyForwards
@@ -123,47 +113,45 @@ pub const MsgBuffer = struct {
     }
 };
 
-fn doTest(nb: *u8) void {
-    nb.* += 1;
-}
-
-test "MsgBuffer" {
+test "Buffer" {
     const Case = struct {
         input: []const u8,
         nb: u8,
     };
-    const alloc = std.testing.allocator;
+
     const cases = [_]Case{
         // simple
-        .{ .input = "2:ok", .nb = 1 },
+        .{ .input = .{ 2, 0 } ++ "ok", .nb = 1 },
         // combined
-        .{ .input = "2:ok3:foo7:bar2:ok", .nb = 3 }, // "bar2:ok" is a message, no need to escape "2:" here
+        .{ .input = .{ 2, 0 } ++ "ok" ++ .{ 3, 0 } ++ "foo", .nb = 2 },
         // multipart
-        .{ .input = "9:multi", .nb = 0 },
+        .{ .input = .{ 9, 0 } ++ "multi", .nb = 0 },
         .{ .input = "part", .nb = 1 },
         // multipart & combined
-        .{ .input = "9:multi", .nb = 0 },
-        .{ .input = "part2:ok", .nb = 2 },
+        .{ .input = .{ 9, 0 } ++ "multi", .nb = 0 },
+        .{ .input = "part" ++ .{ 2, 0 } ++ "ok", .nb = 2 },
         // multipart & combined with other multipart
-        .{ .input = "9:multi", .nb = 0 },
-        .{ .input = "part8:co", .nb = 1 },
+        .{ .input = .{ 9, 0 } ++ "multi", .nb = 0 },
+        .{ .input = "part" ++ .{ 8, 0 } ++ "co", .nb = 1 },
         .{ .input = "mbined", .nb = 1 },
         // several multipart
-        .{ .input = "23:multi", .nb = 0 },
+        .{ .input = .{ 23, 0 } ++ "multi", .nb = 0 },
         .{ .input = "several", .nb = 0 },
         .{ .input = "complex", .nb = 0 },
         .{ .input = "part", .nb = 1 },
         // combined & multipart
-        .{ .input = "2:ok9:multi", .nb = 1 },
+        .{ .input = .{ 2, 0 } ++ "ok" ++ .{ 9, 0 } ++ "multi", .nb = 1 },
         .{ .input = "part", .nb = 1 },
     };
-    var msg_buf = try MsgBuffer.init(alloc, 10);
-    defer msg_buf.deinit(alloc);
+
+    var b: [MaxSize]u8 = undefined;
+    var buf = Buffer{ .buf = &b };
+
     for (cases) |case| {
         var nb: u8 = 0;
-        var input: []const u8 = case.input;
+        var input = case.input;
         while (input.len > 0) {
-            const parts = msg_buf.read(alloc, input) catch |err| {
+            const parts = buf.read(input) catch |err| {
                 if (err == error.MsgMultipart) break; // go to the next case input
                 return err;
             };

src/server.zig

@@ -30,7 +30,8 @@ const CloseError = jsruntime.IO.CloseError;
 const CancelError = jsruntime.IO.CancelError;
 const TimeoutError = jsruntime.IO.TimeoutError;
 
-const MsgBuffer = @import("msg.zig").MsgBuffer;
+const MsgBuffer = @import("msg.zig").Buffer;
+const MaxSize = @import("msg.zig").MaxSize;
 const Browser = @import("browser/browser.zig").Browser;
 const cdp = @import("cdp/cdp.zig");
 
@@ -161,7 +162,7 @@ pub const Ctx = struct {
         // read and execute input
         var input: []const u8 = self.read_buf[0..size];
         while (input.len > 0) {
-            const parts = self.msg_buf.read(self.alloc(), input) catch |err| {
+            const parts = self.msg_buf.read(input) catch |err| {
                 if (err == error.MsgMultipart) {
                     return;
                 } else {
@@ -434,8 +435,8 @@ pub fn handle(
 
     // create buffers
     var read_buf: [BufReadSize]u8 = undefined;
-    var msg_buf = try MsgBuffer.init(loop.alloc, BufReadSize * 256); // 256KB
-    defer msg_buf.deinit(loop.alloc);
+    var buf: [MaxSize]u8 = undefined;
+    var msg_buf = MsgBuffer{ .buf = &buf };
 
     // create I/O completions
     var accept_completion: Completion = undefined;