From sflaniga at redhat.com Wed Dec 14 07:46:59 2016 From: sflaniga at redhat.com (Sean Flanigan) Date: Wed, 14 Dec 2016 17:46:59 +1000 Subject: [zanata-devel] Zanata's XliffReader class Message-ID: Hi Haiwei, You were asking on freenode: Hey guys, I have a question about supporting Xliff type of files. > Inside XliffCommon.java file, the function getContentElementList() defines > several element that are not eligible to use in Xliff type of file. In > XliffReader.java file, get ElementValue() function, it throws errors if > legalElements.contains(localName). I want to ask why do you guys choose to > exclude these tags and what if the input sdlxlf file contains one of these > tag > Thanks a lot in advance! Much appreciate if you can email your > thoughts to my email haiwei.su at motaword.com. The existing XLIFF support in Zanata is very old and unfortunately only supports one particular subset of XLIFF (very flat XLIFF, almost like properties files), not general XLIFF support at all. That's why we don't tend to mention XLIFF support, eg on the zanata.org home page. I think the code you found in XliffReader just tries to give a better error message, depending on whether the elements are legal or illegal in XLIFF. XliffReader was not written to handle these nested elements, even where they are legal. If someone were to implement proper support for XLIFF, it would be better to make a new OkapiFilterAdapter for XLIFF, which should be able to handle hierarchical XLIFF much better than the old approach ever could. In terms of complexity, TSAdapter is probably closest to what an XLIFFAdapter would need to cover. If you have any more questions, please join and ask on the zanata-devel mailing list (CCed): http://zanata.org/mailing-lists/ Regards Sean. -- Sean Flanigan Principal Software Engineer Globalisation Tools Engineering Red Hat -------------- next part -------------- An HTML attachment was scrubbed... URL: