On 2011-02-25, Antoine Levy-Lambert wrote: > Following up on a thread on gump mailing list. [1]
> The change that you mentioned in your earlier email is likely this one [2]: > r1061156 | bodewig | 2011-01-20 03:25:35 -0500 (Thu, 20 Jan 2011) | 1 line > Give build.sysclasspath a chance to influence a forked VM's > bootclasspath even if the user didn't supply one herself Yes > And in fact the gump failure started on Jan 22 [3] > Funny thing about this test taskdefs/import-test.xml : > I get it to fail using ant 1.8.2 (so before the last change that you > did concerning bootclasspath) when I run it like that. > ant -Dbuild.sysclasspath=only -lib lib/optional/ant-antunit-1.1.jar -f > src/tests/antunit/taskdefs/import-test.xml > If I run the test in this manner : > ant -Dantunit.testcase=taskdefs/import-test.xml > -Dbuild.sysclasspath=only -lib lib/optional/ant-antunit-1.1.jar > antunit-tests Another thing that changed around the same time is that Gump passes in build.sysclasspath as a system property now. Sorry I didn't mention it earlier I didn't think of it. In fact I think this change is more likely to affect AntUnit tests than the change to Ant. Stefan --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@ant.apache.org For additional commands, e-mail: dev-h...@ant.apache.org
