Sure is ;-) I've heard good things about JProbe Coverage, and I should probably have mentioned it as a possible candidate. Sorry about that!
On the other hand, I've also heard tools in the JProbe family, although very good, can be quite expensive... If indeed there's a substantial price difference between Clover and JProbe Coverage, it then all depends on the price one puts on dynamic runtime instrumentation... So does anyone know the full price for a single license of each of these tools? And personally, I'd like to know *how* JProbe Coverage does runtime instrumentation... I can see only two ways to do it, with are thru either a custom classloader, and that requires an app that does not break when run under such a custom classloader (!), or thru a custom JVM, which I would be very wary off. Or is there a third way I didn't envision, or is it simply magic??? Thanks, --DD -----Original Message----- From: Gordon Tyler [mailto:gordon.tyler@;sitraka.com] Sent: Friday, November 01, 2002 10:39 AM To: Ant Users List Subject: Re: JUnitReport - more info to display? Dominique Devienne wrote: > A colleague of mine wrote a very smart static analyzer of the test code to > figure out which calls it was making (by introspecting the bytecode), but > even that is not enough to figure out what really gets called at runtime, > which is necessarily a subset of the calls discovered in the bytecode (by > direct calls), and doesn't take into account the interactions of the code > tested itself. It's still a much better metric of the testing done than just > the number and names of the test methods... --DD Although Clover has the disdvantage that it requires pre-instrumentation of your code, in contrast to something JProbe Coverage which does runtime instrumentation. Ciao, Gordon -- Gordon Tyler Software Developer, R&D Sitraka -- Performance is Mission Critical -- To unsubscribe, e-mail: <mailto:ant-user-unsubscribe@;jakarta.apache.org> For additional commands, e-mail: <mailto:ant-user-help@;jakarta.apache.org>
