These test files are intended to test generated text, such as the word "Note" that appears alongside a <note> element.
They also test index processing for each supported language, in most cases by declaring index terms that begin with
every (expected) character in the language. Some languages (such as Korean) only use a subset of possible characters.
The map bookmap.ditamap includes chapters, front matter, and topic content that
will generate NLS text for PDF output.
For each supported language:
- A map sets the
xml:langattribute to that language - Sections of the bookmap are reused by conref
- A topic specific to the language is referenced as a chapter
- Within the topic, index terms are defined or reused by conref
The Ant build file has a variety of targets to test any of the following:
- Build a single language to PDF, XHTML, or HTML5
- Build a single language to PDF, XHTML, and HTML5
- Build PDF, HTML5, or XHTML of all supported languages
- Build PDF, HTML5, and XHTML of all supported languages
- To test an unsupported language by building an unsupported language (Welsh) to any or all of PDF, XHTML, or HTML5
See comments in the build file for instructions.
The following languages (at least) need improved index coverage that includes terms with their own character set. For now, these use index terms taken from localized strings that already appear elsewhere as generated text:
- Japanese
- Simplified Chinese