The TestTrendChart is not correct:
the test statistics of the first available build are reported also for dates before the build date.

consider date_min as the lowest build date among all the builds of all jobs;
if the first available build of a job is at date1 > date_min, for each date in range [date_min, date], the test statistics of build at date1 are added also to the range above.

Instead of having
date1: 1 tests // one job contributes one test
date2: 1 tests // one job contributes one test
date3: 2 tests // two jobs contribute one test each
date4: 3 tests
we have
date1: 2 tests // should be 1: one of this tests don't exist at this date
date2: 2 tests // should be 1: one of this tests don't exist at this date
date3: 2 tests // correct
date4: 3 tests

Attached are the two graphs I have generated. It can be seen that some tests are available only after a certain date, but in the wrong are added also before that date.

Has anybody experienced this problem?