On Wed, 2 Aug 2017, Marek Polacek wrote: > Hmm, how about this, then? > > "operand of ?: changes signedness from %qT to %qT due to unsignedness of > other operand" > > I couldn't come up with anything more brief yet conveying all the information. > I don't like adding "second"/"third"/... very much; we should offer a good > location already. > > Bootstrapped/regtested on x86_64-linux, ok for trunk?
OK. -- Joseph S. Myers jos...@codesourcery.com