From fdc67a08f4a107fdc63a93021379fb3107762e71 Mon Sep 17 00:00:00 2001
From: Mitchell Hashimoto <mitchell.hashimoto@gmail.com>
Date: Tue, 6 Feb 2024 08:45:41 -0800
Subject: [PATCH] terminal: add tests for incomplete utf-8, fix one bug

---
 src/simd/vt.zig         | 17 +++++++++++++++++
 src/terminal/stream.zig | 37 ++++++++++++++++++++++++++++++++++---
 2 files changed, 51 insertions(+), 3 deletions(-)

diff --git a/src/simd/vt.zig b/src/simd/vt.zig
index 2e59957c1..dc1c0a511 100644
--- a/src/simd/vt.zig
+++ b/src/simd/vt.zig
@@ -119,3 +119,20 @@ test "decode invalid UTF-8" {
 
     try testing.expectEqual(@as(u32, 0xFFFD), output[5]);
 }
+
+// This is testing our current behavior so that we know we have to handle
+// this case in terminal/stream.zig. If we change this behavior, we can
+// remove the special handling in terminal/stream.zig.
+test "decode invalid leading byte isn't consumed or replaced" {
+    const testing = std.testing;
+
+    var output: [64]u32 = undefined;
+
+    {
+        const str = "hello\xFF";
+        try testing.expectEqual(DecodeResult{
+            .consumed = 5,
+            .decoded = 5,
+        }, utf8DecodeUntilControlSeq(str, &output));
+    }
+}
diff --git a/src/terminal/stream.zig b/src/terminal/stream.zig
index 70a643bf8..5d016bffc 100644
--- a/src/terminal/stream.zig
+++ b/src/terminal/stream.zig
@@ -71,6 +71,7 @@ pub fn Stream(comptime Handler: type) type {
             // If the scalar UTF-8 decoder was in the middle of processing
             // a code sequence, we continue until it's not.
             while (self.utf8decoder.state != 0) {
+                if (offset >= input.len) return;
                 try self.next(input[offset]);
                 offset += 1;
             } else if (self.parser.state != .ground) {
@@ -106,9 +107,7 @@ pub fn Stream(comptime Handler: type) type {
                 // to the scalar parser.
                 if (input[offset] != 0x1B) {
                     const rem = input[offset..];
-                    for (rem) |c| {
-                        try self.next(c);
-                    }
+                    for (rem) |c| try self.next(c);
                     return;
                 }
 
@@ -1456,6 +1455,38 @@ test "stream: print" {
     try testing.expectEqual(@as(u21, 'x'), s.handler.c.?);
 }
 
+test "simd: print invalid utf-8" {
+    const H = struct {
+        c: ?u21 = 0,
+
+        pub fn print(self: *@This(), c: u21) !void {
+            self.c = c;
+        }
+    };
+
+    var s: Stream(H) = .{ .handler = .{} };
+    try s.nextSlice(&.{0xFF});
+    try testing.expectEqual(@as(u21, 0xFFFD), s.handler.c.?);
+}
+
+test "simd: complete incomplete utf-8" {
+    const H = struct {
+        c: ?u21 = null,
+
+        pub fn print(self: *@This(), c: u21) !void {
+            self.c = c;
+        }
+    };
+
+    var s: Stream(H) = .{ .handler = .{} };
+    try s.nextSlice(&.{0xE0}); // 3 byte
+    try testing.expect(s.handler.c == null);
+    try s.nextSlice(&.{0xA0}); // still incomplete
+    try testing.expect(s.handler.c == null);
+    try s.nextSlice(&.{0x80});
+    try testing.expectEqual(@as(u21, 0x800), s.handler.c.?);
+}
+
 test "stream: cursor right (CUF)" {
     const H = struct {
         amount: u16 = 0,