fix(markup): prevent a multi-byte character after a tag from rendering as broken

When the read buffer ends inside a multi-byte character that directly
follows a tag, keep the cut bytes for the next read instead of emitting
them one by one as broken characters.

Assisted-by: Claude Code:claude-opus-5-5

modules/charset/escape_stream.go

@@ -246,6 +246,9 @@ func (e *htmlChunkReader) readRunes() (parts [][]byte, partInTag []bool, _ error
 			}
 		}
 		if curPartLen == 0 {
+			if e.readErr == nil && curPartEnd == len(e.readBuf) && !utf8.FullRune(curPart) {
+				break // the part is a rune cut by the read boundary, keep it for the next read
+			}
 			// actually it's impossible that the part doesn't contain any valid rune,
 			// the only case is that the cap(readBuf) is too small, or the origin contain indeed doesn't contain any valid rune
 			// * try to leave the last 4 bytes (possible longest utf-8 encoding) to next round

modules/charset/escape_test.go

@@ -234,4 +234,12 @@ func TestHTMLChunkReader(t *testing.T) {
 	testReadChunks(t, 5, input, []textPart{{"<Aé", true}, {"啊", true}, {"🌞>", true}, {"Aé", false}, {"啊", false}, {"🌞", false}})
 	testReadChunks(t, 6, input, []textPart{{"<Aé", true}, {"啊", true}, {"🌞>", true}, {"A", false}, {"é啊", false}, {"🌞", false}})
 	testReadChunks(t, 7, input, []textPart{{"<Aé啊", true}, {"🌞>", true}, {"A", false}, {"é啊", false}, {"🌞", false}})
+
+	// a rune right after a tag must stay whole when the chunk boundary cuts it
+	for _, r := range []string{rune2, rune3, rune4} {
+		for chunkSize := 6; chunkSize < 5+len(r); chunkSize++ {
+			testReadChunks(t, chunkSize, "<abc>"+r, []textPart{{"<abc>", true}, {r, false}})
+		}
+	}
+	testReadChunks(t, 6, "<abc>"+rune3[:2], []textPart{{"<abc>", true}, {rune3[:1], false}, {rune3[1:2], false}})
 }