How did you determine that the coverage was 30% for 1 in the first
place?

   - Oliver

My experience with code coverage is to use a coverage tool. Something like
JProbe can be used to determine what you test cases have covered. You
instrument the application with the coverage tool, run your scripts/jUnit
tests. It will then tell what was executed and how many times. Without
something like JProbe I'm not sure how you would do it.

Hi,

I can suggest the code-coverage tool emma/emmajava. It sometimes is a
bit difficult to setup, but it is worth the setup-trouble, because it
gives you really great code-coverage statistics.

Note however, that code-coverage (measured in percentage of
lines/blocks/methods executed) is not a perfect kind of measurement. e.g.:

String sql="SELECT * FROM "
if(t1) {
  sql+="table1";
}
if(t2) {
  sql+="table2";
}
execute(sql);

If you execute two tests on that code, one with t1==true and t2==false
and one test with t1==false and t2==true, the code is covered to 100%.

But of course that does not mean that the code is correct. If you would
do a third test with t1==false and t2==false, it would fail!

Note that (AFAIK) there does not exist any tool that can sensefully deal
with this problem.

Ciao,
Ingo