I could see some use to that, my main concern would be as xml2brl is a document processing tool getting the correct output might be harder than for liblouis. What I mean is that some of the layout may be able to differ and still be valid. I think it would need suggestions on how to use it, eg. should I do a translation with xml2brl and hand correct the output file to create the test output?
Michael Whapples On 09/11/09 11:17, Christian Egli wrote:
On Sat, 2009-11-07 at 21:40 +0000, Michael Whapples wrote:Hello, I said in my previous message I was sure that there were some issues with liblouisxml I had noticed, well here they are:Thanks Michael for this testing. What I would like to grow out of this is some simple test framework that just takes a bunch of input files, passes them to xml2brl with a set of options and compares the output to a set of defined responses. In your case you define an xml input file and you also have a brl output that you expect given a set of configuration settings. This could be done in a similar way to what I did for liblouis where you place a file into a certain directory and the two columns in this file are automatically tested in the 'make check' process (http://code.google.com/p/liblouis/source/browse/trunk/tests/table_test_corpuses/README). I could see a test directory where you place your test file. Each directory could contain an input file, an output file (the expected output) and probably a config file. Then a simple Perl script just invokes xml2brl with the right config file and compares the output with the given one. I can hack up this Perl script if you think this is a valid idea (and contribute some test data :-)). Thanks Christian
For a description of the software and to download it go to http://www.jjb-software.com