Few of our tests (Chasrset and Symbols) do wrong assumption, I was looking at this when I was on the DecimalFormat and the Collator but never came up with a fix sorry…

The jtreg results would be interesting also, anyone with results?

Mario

Posted the output of a run at http://www.klomp.org/mark/classpath/icedtea-1.0-mauve-run.out

Since mauve doesn’t have an easy option not to run any awt/swing related tests this is just done by hand as follows:

PATH=~/src/icedtea/openjdk/control/build/linux-i586/bin:$PATH ~/src/icedtea/openjdk/control/build/linux-i586/bin/java Harness -vm ~/src/icedtea/openjdk/control/build/linux-i586/bin/java -timeout 30000 -showpasses java.io java.net java.lang java.math java.security java.text java.util java.nio java.rmi java.sql javax.crypto javax.accessibility javax.activity javax.management javax.net javax.rmi javax.xml javax.transaction javax.sql javax.security javax.naming

Note that this is the first time that I have run the Mauve tests on anything not GNU Classpath derived and some tests fail because of different assumptions of what the core classes should do and some fail because they have hard coded GNU dependencies (bad us!). So some cleaning up of the tests is in order.

Can you point to the Mauve results? We should run jtreg on IcedTea as well. This is really awesome work built on all the cumulative effort of Classpath & Friends! I especially want to thank
Andrew Haley, Thomas Fitzsimmons, Lillian Angel, Tania Bento,
Kyle Galloway, and Francis Kung for helping show the way to Freedom.

Regards,

–Tom