Przeglądaj źródła

LibWeb: Fix utf16-be check in HTMLEncodingDetection

The utf-16be check mistakenly skipped index 3, so was not checking the
correct bytes. This meant UTF16-BE files could fail to decode.
MacDue 1 rok temu
rodzic
commit
fc41c282ec

+ 20 - 0
Tests/LibWeb/Layout/expected/utf-16-be-xhtml-file-should-decode-correctly.txt

@@ -0,0 +1,20 @@
+Viewport <#document> at (0,0) content-size 800x600 children: not-inline
+  BlockContainer <html> at (0,0) content-size 800x41.46875 [BFC] children: not-inline
+    BlockContainer <body> at (8,16) content-size 784x17.46875 children: not-inline
+      BlockContainer <(anonymous)> at (8,16) content-size 784x0 children: inline
+        TextNode <#text>
+      BlockContainer <p> at (8,16) content-size 784x17.46875 children: inline
+        line 0 width: 29.21875, height: 17.46875, bottom: 17.46875, baseline: 13.53125
+          frag 0 from TextNode start: 1, length: 15, rect: [8,16 29.21875x17.46875]
+            "好啦朋友們"
+        TextNode <#text>
+      BlockContainer <(anonymous)> at (8,49.46875) content-size 784x0 children: inline
+        TextNode <#text>
+
+ViewportPaintable (Viewport<#document>) [0,0 800x600]
+  PaintableWithLines (BlockContainer<HTML>) [0,0 800x41.46875] overflow: [0,0 800x49.46875]
+    PaintableWithLines (BlockContainer<BODY>) [8,16 784x17.46875] overflow: [8,16 784x33.46875]
+      PaintableWithLines (BlockContainer(anonymous)) [8,16 784x0]
+      PaintableWithLines (BlockContainer<P>) [8,16 784x17.46875]
+        TextPaintable (TextNode<#text>)
+      PaintableWithLines (BlockContainer(anonymous)) [8,49.46875 784x0]

BIN
Tests/LibWeb/Layout/input/utf-16-be-xhtml-file-should-decode-correctly.html


+ 4 - 2
Userland/Libraries/LibWeb/HTML/Parser/HTMLEncodingDetection.cpp

@@ -251,10 +251,12 @@ Optional<ByteString> run_prescan_byte_stream_algorithm(DOM::Document& document,
     // https://html.spec.whatwg.org/multipage/parsing.html#prescan-a-byte-stream-to-determine-its-encoding
 
     // Detects '<?x'
-    if (!prescan_should_abort(input, 6)) {
+    if (!prescan_should_abort(input, 5)) {
+        // A sequence of bytes starting with: 0x3C, 0x0, 0x3F, 0x0, 0x78, 0x0
         if (input[0] == 0x3C && input[1] == 0x00 && input[2] == 0x3F && input[3] == 0x00 && input[4] == 0x78 && input[5] == 0x00)
             return "utf-16le";
-        if (input[0] == 0x00 && input[1] == 0x3C && input[2] == 0x00 && input[4] == 0x3F && input[5] == 0x00 && input[6] == 0x78)
+        // A sequence of bytes starting with: 0x0, 0x3C, 0x0, 0x3F, 0x0, 0x78
+        if (input[0] == 0x00 && input[1] == 0x3C && input[2] == 0x00 && input[3] == 0x3F && input[4] == 0x00 && input[5] == 0x78)
             return "utf-16be";
     }