XSLT Tests

Getting the XSLT test suite

1. Pick a directory under which the tests will live. From now on it will be referred to as $TESTHOME. Note that the only thing created in $TESTHOME will be an xalan-test/ directory.
2. Use the SVN repository base URL http://svn.apache.org/repos/asf/
3. Check out the trunk of the xalan/test module.

For example, on a Unix system you might do

 cd ~
 svn co http://svn.apache.org/repos/asf/xalan/test/trunk xalan-test

Running the tests

1. Compile Seamonkey with the --enable-tests option. This is needed to run the tests until the user interface is ported to Firefox.
2. From now on, we will refer to ${TESTHOME}/xml-xalan/test/tests (this is the directory that has conf and conf-gold as subdirectories) as $TESTDIR. We will refer to your Mozilla source directory as $MOZDIR.
3. Go to $TESTDIR and run $MOZDIR/content/xslt/tests/buster/helper/generate-rdf.pl. This will generate a file called xalan.rdf in $TESTDIR. Steps 1-3 might look something like:

   cd ~/xml-xalan/test/tests
   perl ~/mozilla/debug/mozilla/content/xslt/tests/buster/helper/generate-rdf.pl
   

4. Start a build that does NOT have the changes being tested in it. Go to the Tools > Web Development > Xalan Tests menu option. This opens a "Xalan testcase harness" window. Resize this window to be big enough to show at least three lines of the tree at the bottom of the window.
5. Click the "Browse" button and browse for the xalan.rdf file that you created in step 3.
6. Select the "conf" item in the tree at the bottom of the "Xalan testcase harness" window.
7. Click the button labeled "run checked tests"
8. Wait until the test completes.
9. Go to File > Save Results and save the results to a file of your choice.
10. Repeat step 4 with a build that DOES have the changes being tested in it.
11. Go to File > Import results and import the results you saved in step 8.
12. Repeat steps 5-8.
13. Compare the results -- the "Success" column will have the results with the changes being tested, while the "Previously" column is the baseline.