Hi all I'm currently reviewing the big change that introduced the regression in 1.10.2 as we may have overlooked more than this issue. It is a *big* change and so it is taking time.
While looking at the hunk starting at https://github.com/apache/ant/commit/b7d1e9bde44cb8e5233d6e70bb96e14cbb2f3e2d#diff-3cabe19d89e908d993d999100d888b6eL256 I think the buildup of the path is now backwards. The original code would add all elements of dependClasspath to the path that are not members of destPath to the new Path p. Unless I am mistaken the Difference created in the new code will in addition retain all elements of destPath that are not in dependClasspath. So we get all elements that are in exactly one of the two - which is more than we would have with the old code unless destPath is a subset of dependClasspath. Does anybody else read it the same way? Or can tell me I'm wrong? Unfortunately our test coverage for the depend task is so close to non-existent that we can't tell from the tests whether we've broken anything. Stefan --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@ant.apache.org For additional commands, e-mail: dev-h...@ant.apache.org
