When this thread first came up, the point was that the package maintainer should have the option to not compile with -g.
That was fine. However, the final policy depreciates the current practice in favor of a new and almost unknown mechanism. That's not so good: The way I see it, this whole issue is about optimizing compile time. The proposed solution gains some speed but specifies that the the maintainer instrument compilation so that debian/rules can say whether or not to build with -g. In many packages, this will be easy. In some packages, this will be irrelevant. However, in some packages it will be rather difficult. So: I object to depreciating the current -g + strip mechanism. I believe that we should allow package maintainers to make the decision between compilation speed and simplicity. [I'd be less touchy about this if I felt that most developers were aware of this new mechanism and that they all approved of it.] Thanks, -- Raul
