Hello David N Bertoni, Tuesday, April 16, 2002, 5:59:13 PM, David N Bertoni/Cambridge/IBM wrote:
DNBCI> Maybe it doesn't matter to you, but it matters according to the to the XML DNBCI> recommendation. It also matters to those who wants conforming tools. Yes.. I also want conforming tools - but sometimes it's nice to enable some dirty tricks ;-) I don't want Xalan to ignore invalid characters generally... but perhaps using a parameter to make Xalan ignore these characters. DNBCI> Any document which contains invalid characters is not really an XML DNBCI> document. Either fix your document, or don't try to process it using using DNBCI> XML tools. Nice hint.. and of course I try to fix invalid characters. But there are two cases where it's really annoying: 1) JDK 1.4 logger (using XML format) sometimes can sometimes uses the "wrong" character set producng invalid characters... I know it's possible to set the encoding for the logging, but sometimes others use my XSLT... 2) German umlauts cause problems when using UTF-8 and Windows, that is using a Windows editor inserting umlauts. In my case (which caused me writing the first mail), I commited an XSLT to CVS, adding a comment including an umlaut - and the XSLT didn't work anymore. Of course I try to fix invalid characters, but since it's possible with "pure" Java to read these corrupt files without any error, it would be nice if Xalan could inherit this "feature" if I want to. Best regards, Jens
