(2010/06/06 20:27), Yusuke ENDOH wrote:
> Any code modification may often make a new bug even if it looks like
> trivial one.  We should select a solution with no code modification
> if any, especially during release process.

You are speaking too generally.  Where's the bug you are talking about?

> By adding a note to README, we can explicitly state that ext/readline
> cannot be linked with readline6.  Why do you think it is not enough?

Because people can ignore such statements.  Machine checking is stronger than a
mare readme.

But OK, people's ignoring license agreements is definitely not our fault.  If
you think we should not care about that situation, I can accept that.  Though I
don't think the status quo is a problem.