On Tue, 4 Feb 2014, Marek Polacek wrote: > Starting with revision 206620, if we have LHS op= RHS and RHS has > side effects, we preevaluate RHS first, which means that we wrap it > in C_MAYBE_CONST_EXPR_EXPR + SAVE_EXPR. This regressed -Wconversion, > since we now issue bogus warnings. The reason is that conversion_warning > doesn't grok C_MAYBE_CONST_EXPR_EXPR/SAVE_EXPR, meaning that it > just calls unsafe_conversion_p on it that warns. But for e.g. relational > and equality operators that's useless as those always yield 0 or 1. > This patch fixes this by unwrapping the expr first. > > Regtested/bootstrapped on x86_64-linux, ok for trunk?
OK. -- Joseph S. Myers jos...@codesourcery.com
