ditto that will break a lot of stuff

Still think it's worth it. I adds so much noise to Mojos (especially reports) that this is a great fix.

Who is actually using PU outside of Codehaus or Maven space?

Who is actually using PU outside of Codehaus or Maven space?

I don't really think it matters. If it goes inside a major release, users will have to adjust, but the change makes sense to me.

FWIW, those have to be reflected in maven-utils as well.

Who is actually using PU outside of Codehaus or Maven space?

I don't really think it matters. If it goes inside a major release, users will have to adjust, but the change makes sense to me.

in this case, I was using it for the taglist-maven-plugin. Someone was using it in canGenerateReport(). This method can neither throw a MojoExecutionException nor a MavenReportException, which is problematic:

mojohaus/taglist-maven-plugin@dac102d#diff-d78dd9fa5c0bfb4f18b93522e585ad3e35dc9eb0315763dd26e7ea7465d04764R466-R473

For this reason, I created yet another issue: https://issues.apache.org/jira/browse/MNG-7393

I do not think throwing a RuntimeException is particularly helpful here, though.

Question: When the code will move to NIO 2, will IOExceptions come back?

Question: When the code will move to NIO 2, will IOExceptions come back?

Good point. Not at PC today, but from memory: yes, probably.

Question: When the code will move to NIO 2, will IOExceptions come back?

Good point. Not at PC today, but from memory: yes, probably.

If this evaluates to true, then this PR should be dropped.

@bmarwell Do you still want to hold onto this PR although I have a provided a reason not to?

@michael-o Good question. I initially opened this issue because of how the maven report works. But we can deal with it later as well.