On Mon, 7 Sep 2020 at 23:53, Bruce Korb via Gcc <gcc@gcc.gnu.org> wrote: > > On Mon, Sep 7, 2020 at 3:45 PM Florian Weimer <f...@deneb.enyo.de> wrote: > > > > * Bruce Korb via Gcc: > > > > > I don't write a lot of code anymore, but this sure seems like a > > > gratuitous irritation to me. I've been using > > > > > > // FALLTHRU and > > > // FALLTHROUGH > > > > > > for *DECADES*, so it's pretty incomprehensible why the compiler should > > > have to invalidate my code because it thinks a different coding > > > comment is better. > > > > It's not clear what you are talking about. > > > > Presumably you placed the comment before a closing brace, and not > > immediately before the subsequent case label? > > Nope. I had /* FALLTHROUGH */ on the line before a blank line before > the case label. After Googling, I found an explicit reference that you > had to spell it: // fall through > I did that, and it worked. So I'm moving on, but still ...
The canonical reference is https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html#index-Wimplicit-fallthrough and it says FALLTHROUGH is fine (except with -Wimplicit-fallthrough=5 which "doesn’t recognize any comments as fallthrough comments, only attributes disable the warning").
