Stuart, I'm a *huge* fan of your Japi diffs (I admit that I have a sort of obsessive compulsion about it, getting happy with every small percentage increase and sad when it decreases).
That said, I have a few questions to ask, see below. Stuart Ballard wrote: > Japi diff jdk15 vs harmony: > Full results: > http://www.kaffe.org/~stuart/japi/htmlout/h-jdk15-harmony.html > > Changes since last run: > > -Comparison run at Thu Nov 2 13:05:30 2006 GMT > -jdk15 API scanned at 2006/11/02 05:19:25 EST > -harmony API scanned at 2006/11/02 06:15:37 GMT > +Comparison run at Fri Nov 3 13:06:47 2006 GMT > +jdk15 API scanned at 2006/11/03 05:19:33 EST > +harmony API scanned at 2006/11/03 11:15:31 GMT > -javax.rmi.ssl: 100% missing > +javax.rmi.ssl: 100% good > -javax.swing.plaf.metal: 98.8% good, 1.19% missing > +javax.swing.plaf.metal: 99.98% good, 0.01% missing > -Total: 94.59% good, 0.06% minor, 0.03% bad, 5.3% missing, 0.02% abs.add > +Total: 94.66% good, 0.06% minor, 0.03% bad, 5.23% missing, 0.02% abs.add > -Packages: 11 missing. > -Classes: 41 minor, 2 bad, 53 missing. > +Packages: 10 missing. > +Classes: 41 minor, 2 bad, 52 missing. > -javax.rmi.ssl: > -Missing > -package javax.rmi.ssl: missing in harmony > - > -class javax.swing.plaf.metal.OceanTheme: missing in harmony It tells me there are 10 packages and 52 classes missing. Great. Then it gives me a percentage. Who is that percentage calculated? Is it by class coverage? method coverage? package coverage? a weighted sum of all the above? > Japi diff harmony vs jdk15: > Full results: > http://www.kaffe.org/~stuart/japi/htmlout/h-harmony-jdk15.html Ok, now it's the reverse... > Changes since last run: > > -Comparison run at Thu Nov 2 13:08:18 2006 GMT > -harmony API scanned at 2006/11/02 06:15:37 GMT > -jdk15 API scanned at 2006/11/02 05:19:25 EST > +Comparison run at Fri Nov 3 13:09:50 2006 GMT > +harmony API scanned at 2006/11/03 11:15:31 GMT > +jdk15 API scanned at 2006/11/03 05:19:33 EST > +javax.rmi.ssl: 100% good > -javax.swing.plaf.metal: 99.34% good, 0.62% minor, 0.03% missing > > -Total: 90.88% good, 0.28% minor, 0.13% bad, 8.69% missing, 0% abs.add > +Total: 90.89% good, 0.28% minor, 0.13% bad, 8.69% missing, 0% abs.add How is it possible that jdk15 vs harmony is 94.66% and harmony vs jdk15 is 90.88% Also in jdk vs harmony we get that javax.swing.plaf.metal: 99.98% good, 0.01% missing while in harmony vs jdk we get that javax.swing.plaf.metal: 99.34% good, 0.61% minor, 0.03% missing Close, clearly, but why should they be different? Thanks in advance for any insight. Ah, also, would it be possible to have a list of the percentage over time? It would be very interesting to plot the evolution of coverage over time. -- Stefano.