Overriding built-in parser for TikaCLI with Tika 1.9

[Stephan Mühlstrasser]

Hi,

I have an existing example how to override a built-in parser with the 
ServiceRegistry mechanism. The file 
"META-INF/services/org.apache.tika.parser.Parser" lists my desired 
parser class, and the directory containing the META-INF directory is 
added to the class path.

With Tika 1.0 this worked fine for both TikaCLI and direct use of the 
Tika API.

Now I tested with Tika 1.9, and the TikaCLI class seems to ignore the 
ServiceRegistry mechanism. The only way I could get TikaCLI to use my 
externally supplied parser was by creating a Tika XML configuration 
file, and by specifying that one with the "--config" option.

Is that intended behavior for TikaCLI now? It looks like the "--config" 
options is fairly new.

And is there any documentation on the syntax of the Tika XML 
configuration file? I was able to find some examples of configuration 
files that I used as blueprints, but I could not find a description of 
the XML syntax on the Tika website.

Thanks
Stephan