Thank you for the reply. Indeed, it is not a very big problem, because one can always choose UTF-8 for the encoding. But files in UTF-8 that consist mostly from national characters require twice more disk space than those in native encoding, and are opened in web browser little slower. I use Xalan to generate large reports, and these aspects are important to me. I understood, that the problem is hard to be resolved in near future, and I have no choice but to use UTF-8.
----- Original Message ----- From: [EMAIL PROTECTED] To: [email protected] Date: Tue, 1 Feb 2005 08:59:29 -0800 Subject: Is this a bug? (numbered entity references in produced HTML) >> Problem: HTML formatter writes numbered entity references >> instead of characters in output encoding (specified in >> "xsl:output" tag), despite the fact that output encoding >> supports these characters. > It shouldn't be a problem, because any web browser should render the > character correctly. >> Formatter writes numbered entity reference for the >> character if the character is greater than the maximum >> character for the encoding (m_maxCharacter), which is >> always 0x7F for non-standard encodings >> (XalanTranscodingServices::getMaximumCharacterValue). >> This makes HTML documents incredibly large when custom >> encoding, added with XMLTransService::addEncoding() is >> used. Produced documents contain numbered entity reference >> for every locale-specific character, because they all have >> codes >0x007Fu. > Yes, this is a known problem with the design of the serializers. I > started working on this about a year ago, but it has not been a high > priority, because very few people have complained about it. You might > want to choose UTF-8 as the output encoding, if the size of the generated > files is too big. Technically, XSLT processors are only required to > support UTF-8 and UTF-16, and fixing this has a potentially significant > performance impact on serialization, because it requires we lookup every > character to determine if the target encoding can represent it. >> If this is a bug, can somebody register it? I couldn't do >> this through JIRA web interface. > As long as you're registered, you should be able to create a bug report. > Dave --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
