g++.old-deja/g++.other/anon7.C

[Richard Kenner]

    Certainly, you can use any means you want to fix it, including getting
    some one else to fix it.  Heck, you can hire someone to fix it for you
    if you want.  As long as it gets fixed.

I think that too is stronger than intended.  We have to balance two things
here: on the one hand, we don't want the system broken, but on the other, we
don't want to discourage contributors by telling them they might end up
responsible for something they shouldn't be and don't know anything about.

If somebody spends time learning, say, the Java front end, and contributes
some nice improvement, but it so happens that it exposes a latent bug in
reload for some target different than the one they were using, can we really
tell them "Sorry, we can't accept the work you've done unless you spend even
more time learning reload" (they might not even have learned about RTL)?
That's quite unreasonable and doing such a thing will discourage
contributions to GCC.

In such a situation, I think it's the obligation of those of us who are
familiar with most of the system to work *with* the person who submitted the
patch that exposed the latent bug.  I don't believe that telling them to go
away is the right approach.