InputStreamReader probably doesn't handle the BOM correctly. But when
you let Xerces do it, it will start with an InputStream, read the bytes
and do the right thing.
Bob
Paul Ekeland wrote:
Hello all,
for some reason yet hard to explain, the problem came from the IOfile. I
was passing an InputStreamReader to the parser, so when I let him deal
with the file by just giving him the appropriate info to find it
everything worked fine: no more "content not allowed in prolog"! If you
know what could have caused it, let me know otherwise I just wanted to
share this (good!) info with you.
Thank you all for your help,
Paul
Andy Clark wrote:
Martin Vysny wrote:
I had the same problem aswell. When you try to save file in
notepad.exe as UTF-8, it places 3-byte invisible UTF-8 character at
the start of xml file. That is causing that goddamn "Content is not
allowed in prolog" message.
That's probably not the problem because Xerces has a custom
UTF-8 reader that knows how to skip the BOM. Unless, of
course, the application doesn't give the parser the chance
to pick the proper java.io.Reader for the input. This can
happen when the application constructs an input source with
a Reader object instead of an InputStream. For example:
Reader reader = new InputStreamReader(stream);
InputSource source = new InputSource(reader);
In this case, the input stream reader will use the default
system encoding, usually ISO Latin 1 on English systems. This
is normally ok because every byte (even with the high bit on)
is valid in that encoding. All except for the UTF-8 byte
order mark which ends up looking like "content [that] is
not allowed in [the] prolog". Even constructing an input
stream reader with the encoding set to "UTF-8" doesn't
help because that will use the Java UTF-8 reader which
doesn't understand the BOM.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
