Error when loading valid Windows-1252 document "System error: -2146697210"

Somehow, sometimes the code below generates an error when loading valid Windows-1252 XML.

It fails on Windows XP Professional x86 SP3 using MSXML6.
It succeeds on Windows 7 Ultimate x64 SP1 using MSXML6.

Note: the code below is written in Delphi, but equivalent code also fails in other environments.

procedure TXMLEOSErrorTestCase.Test;
var
  XmlDocument: IXMLDOMDocument3;
  XmlFileName: string;
begin
  XmlDocument := CoFreeThreadedDOMDocument60.Create();
  XmlFileName :=  TPath.Combine(TPath.GetDirectoryName(ParamStr(0)), '1-Normal.xml');
  if not XmlDocument.load(XmlFileName) then
    Parse(XmlDocument.parseError);
end;

This error occurs during the XmlDocument.load method:

reason: System error: -2146697210.
errorCode: -2146697210
url: C:\temp\1-Normal.xml

I trimmed the XML down to the XML found below.

This is the hex dump of the XML file:

000000: 3C 3F 78 6D 6C 20 76 65  72 73 69 6F 6E 20 3D 20 <?xml version =
000010: 22 31 2E 30 22 20 65 6E  63 6F 64 69 6E 67 3D 22 "1.0" encoding="
000020: 57 69 6E 64 6F 77 73 2D  31 32 35 32 22 3F 3E 3C Windows-1252"?><
000030: 52 4F 57 20 43 69 74 79  3D 22 E0 22 2F 3E 0D 0A ROW City="."/>..

This is the XML:

<?xml version = "1.0" encoding="Windows-1252"?><ROW City="à"/>

Why does the error occur?

(The XML loads perfectly fine in .NET and other environments not using MSXML6, it also works fine on Windows 7 Ultimate x64 SP1).

--jeroen

The behaviour depends on which version of the MSXML6.DLL you have installed.

To reproduce this better, I created another file abnormal.xml, in addition to the normal.xml from the question.

File dump abnormal.xml:

000000: 3C 3F 78 6D 6C 20 76 65  72 73 69 6F 6E 3D 22 31 <?xml version="1
000010: 2E 30 22 20 73 74 61 6E  64 61 6C 6F 6E 65 3D 22 .0" standalone="
000020: 79 65 73 22 3F 3E 3C 52  4F 57 20 43 69 74 79 3D yes"?><ROW City=
000030: 22 E0 22 2F 3E 0D 0A                             "."/>..

File abnormal.xml:

<?xml version="1.0" standalone="yes"?><ROW City="à"/>

File dump normal.xml:

000000: 3C 3F 78 6D 6C 20 76 65  72 73 69 6F 6E 20 3D 20 <?xml version =
000010: 22 31 2E 30 22 20 65 6E  63 6F 64 69 6E 67 3D 22 "1.0" encoding="
000020: 57 69 6E 64 6F 77 73 2D  31 32 35 32 22 3F 3E 3C Windows-1252"?><
000030: 52 4F 57 20 43 69 74 79  3D 22 E0 22 2F 3E 0D 0A ROW City="."/>..

File normal.xml:

<?xml version = "1.0" encoding="Windows-1252"?><ROW City="à"/>

The behaviour I expect is that:

1. abnormal.xml fails, because it does not specify an encoding, but contains a character with the high-bit set
2. normal.xml succeeds, as it conains a single-byte encoding supporting high-bit characters, so characters with high-bit set are allowed

These are the observed scenarios:

MSXML6 FAILURE:

reason: System error: -2146697210.
errorCode: -2146697210
url: file:///C:/My%20Dropbox/XMLEOSErrorTest/Abnormal.xml

reason: System error: -2146697210.
errorCode: -2146697210
url: file:///C:/My%20Dropbox/XMLEOSErrorTest/Normal.xml

MSXML6 SUCCESS:

reason: An invalid character was found in text content.

errorCode: -1072896760
url: file:///C:/My%20Dropbox/XMLEOSErrorTest/Abnormal.xml
srcText: <?xml version="1.0" standalone="yes"?><ROW City="
line: 1
linepos: 50
filepos: 49

This is an overview of what versions fail.
The names of the DLL's between parentheses are from their version information.

failure; XP Professional SP3:
msxml6.dll       version 6.20.1099.0  (MSXML 6.0 SP2)
msxml6r.dll      version 6.0.3883.0   (XML Resources)

success; Windows 7 Ultimate x64 SP1:
msxml6.dll       version 6.30.7600.16385  (MSXML 6.0 SP3)
msxml6r.dll      version 6.30.7600.16385
msxml6r.dll.mui  version 6.30.7600.16385

success; XP Professional SP3:
msxml6.dll       version 6.20.1103.0  (MSXML 6.0 SP3)
msxml6r.dll      version 6.0.3883.0   (XML Resources)

Observations:

• Automatic Windows Update and Microsoft Update will not update MSXML6.DLL to the latest version on Windows XP SP3.
• There is no MSXML6 SP3 for Windows XP.
• The latest version is in fact 6.20.2003.0 for Windows XP SP2 and can be obtained at http://support.microsoft.com/kb/973686 (it does not install on Windows XP SP3).
• The 6.20.1103.0 version of MSXML6.DLL for Windows XP SP3 can be obtained at http://support.microsoft.com/kb/973687 (direct download for XP SP3)
• Microsoft has very little information on their site about this: http://www.google.com/search?q=msxml6+%226.20.1099.0%22+%226.20.1103.0%22+site:microsoft.com
• Searching only for 6.20.1103.0 does not reveal much more: http://www.google.com/search?sourceid=chrome&ie=UTF-8&q=msxml6+%226.20.1103.0%22+site:microsoft.com
• This gives the best search results: http://www.google.com/search?sourceid=chrome&ie=UTF-8&q=msxml6+%226.20.1103.0%22#sclient=psy&num=10&hl=en&q=msxml6+%226.20.1099.0%22+%226.20.1103.0%22

So: when doing MSXML6 work, first put in a check that you indeed have the latest MSXML6.DLL for your target Windows version.

--jeroen