SAX & UTF-8 problem

[Chris]

My SAX parser is choking on UTF-8 encoded files (a "document root element is
missing" error). The problem is three bytes that appear at the beginning of
each file:

0xEF 0xBB 0xBF

If I delete the bytes the problem goes away.

I'm accessing the file by using a FileInputStream and then wrapping it in a
SAX InputSource. My guess is that the InputSource is converting bytes to
chars using the platform's default encoding, rather than UTF-8.

Is there any existing InputSource class or Reader class that will
automatically detect UTF-8 and encode chars correctly? Or do I have to write
my own Reader class to do it?

 

[Roedy Green]

Is there any existing InputSource class or Reader class that will
automatically detect UTF-8 and encode chars correctly? Or do I have to write
my own Reader class to do it?

The thing with most chance of success is a Reader with an explicit
UTF-8 encoding.

Hopefully it will just discard the signature.

 

[Van Ly]

My SAX parser is choking on UTF-8 encoded files (a "document root
element is missing" error). The problem is three bytes that appear at
the beginning of each file:

0xEF 0xBB 0xBF

If I delete the bytes the problem goes away.

I'm getting the same problem with Javadoc. With the three bytes above
(referred to as the BOM) of a UTF-8 file, Javadoc will choke because
it considers the three bytes illegal characters. Of course, removing
the three bytes will get Javadoc going again.

I've tried the "-encoding UTF8" and "-encoding UTF-8" options of Javadoc.
But it still bombs. Anyone reading this can help?

Thanks,
Van

 

[Thomas Weidenfeller]

I've tried the "-encoding UTF8" and "-encoding UTF-8" options of Javadoc.
But it still bombs. Anyone reading this can help?

Write a pre-processor to filter the BOM mark out.

If you have control over the source code (I know, you don't have it for
JavaDoc), see

http://groups.google.com/[email protected]

/Thomas

 

[Thomas Weidenfeller]

Is there any existing InputSource class or Reader class that will
automatically detect UTF-8 and encode chars correctly? Or do I have to write
my own Reader class to do it?

We wrote one a year or so ago:

http://groups.google.com/[email protected]

/Thomas

 

[Chris]

Is there any existing InputSource class or Reader class that will

We wrote one a year or so ago:

http://groups.google.com/[email protected]

Thanks. That worked perfectly.