Finn Bock wrote:
So how do you folks verify that the changes you make does not cause regressions?

Well, the tests never were really effective and weren't updated for changes in head. Have a look at test/java/org/apache/fop/ for a JUnit based approach which hopefully will supercede the batch tests. There is only one test really there, test for other features will follow as time permits. Note that the FO is included as strings in the Java test case by design. I'd like to try to provide a few strings for standard layout master sets, related static content and such which can be easily reused by different test cases, until the static data limit hits and a new test class has to be started.


Reply via email to