Okay. Thanks Dave! Urmi
-----Original Message----- From: David N Bertoni/Cambridge/IBM [mailto:[EMAIL PROTECTED]] Sent: Thursday, April 18, 2002 5:48 PM To: [EMAIL PROTECTED] Subject: RE: Creating an XSLTInputSource for Unicode byte stream Hi Urmi, If you have the choice, always go with the byte stream, because the parser only consumes bytes. Xalan has no feature to break up the bytes, because that would require knowing about the "endian-ness" of the stream -- something we'ed rather not have to worry about. You probably don't want to worry about that either, so I'd recommend you avoid std::wistream. If you decide to go with std::wistream, you'll need to write an adapter between std::wistream and std::istream that knows about the endian-ness of the stream, and breaks each wide character into two bytes, feeding the bytes back in the correct order. Remember you can have both big-endian and little-endian streams being parsed on big-endain or little-endian hardware. Lastly, if you're really developing a "Unicode-compliant" application (I don't know what you really mean by this, but I'm making a few guesses), you'll probably want to support UTF-8 as well, which would preclude using wide character streams. Dave "Chatterjee, Urmi" To: [EMAIL PROTECTED] <URMI.CHATTERJEE cc: (bcc: David N Bertoni/Cambridge/IBM) @ca.com> Subject: RE: Creating an XSLTInputSource for Unicode byte stream 04/18/2002 02:22 PM Please respond to xalan-dev Hi Dave, I am sorry, I did mean an instance of std::wistream. And yes, I meant Unicode characters encoded in UTF-16. Actually, the application that I have to develop is to be Unicode compliant. At this point, I do have the option of choosing between using a byte stream or using a wide character stream as input. Based on what you said, it would look like using a byte stream (beginning with the correct byte order mark, and with the correct encoding declaration) would seem like a quicker option. However, if I decide to go with an std::wistream instance, is there some feature in Xalan which would aid me in breaking it up into bytes? Thanks, Urmi -----Original Message----- From: David N Bertoni/Cambridge/IBM [mailto:[EMAIL PROTECTED]] Sent: Thursday, April 18, 2002 4:32 PM To: [EMAIL PROTECTED] Subject: Re: Creating an XSLTInputSource for Unicode byte stream Hi Urmi, I have no idea what a _wstream* is. Did you mean an instance of std::wistream? A byte stream cannot be an instance of a wstream, since the underlying unit is a wide character. First, make sure your stream starts with the correct byte-order mark, the document has the correct encoding declaration, and initialize a strstream instance with a pointer to the beginning of the stream and the length _in bytes_ of the stream. If you're being provided with an instance of std::wistream, you'll need to provide something which breaks down wide characters into bytes. The parser consumes only bytes. By the way, all this assumes that when you say "Unicode byte stream" you mean Unicode characters encoded in UTF-16. Is that correct? Dave "Chatterjee, Urmi" To: [EMAIL PROTECTED] <URMI.CHATTERJEE cc: (bcc: David N Bertoni/Cambridge/IBM) @ca.com> Subject: Creating an XSLTInputSource for Unicode byte stream 04/18/2002 11:12 AM Please respond to xalan-dev Hi, I have just recently started using Xalan C++ (ver 1.3). My problem is this. I need to transform xml data using XalanTransformer::transform(), where my input is a unicode byte stream in memory of type _wstream *, and the xsl byte stream is also a _wstream * type. However, I do not see an XSLTInputSource constructor for the same. Can you please help me out on the best way to achieve the transformation. Do I need to use a transcoder of some sort? I have already built an ICU enabled build of Xalan. Thanks in advance, Urmi
