Oops, if it is not a bug, please close the report http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57258
Thanks, Mingjie 2013/11/8 Richard Biener <richard.guent...@gmail.com>: > On Fri, Nov 8, 2013 at 10:39 AM, Mike Stump <mikest...@comcast.net> wrote: >> >> On Nov 8, 2013, at 1:32 AM, Bin.Cheng <amker.ch...@gmail.com> wrote: >> >>> On Fri, Nov 8, 2013 at 5:03 PM, Mike Stump <mikest...@comcast.net> wrote: >>>> On Nov 7, 2013, at 5:13 PM, Mingjie Xing <mingjie.x...@gmail.com> wrote: >>>>> Well, it is my understanding that the warning should be emitted for a >>>>> volatile variable only if it is not accessed. Initialization means >>>>> accessing, even though it is not used anywhere. >>>> >>>> Let me try. A warning is useful, if there is no way a conforming program >>>> can tell that the variable exists or not. So, the question is, how can >>>> you notice the variable? Answer, there is no way, so, there is no utility >>>> in having the variable. The warning is to tell the user to remove the >>>> dead variable. >>> >>> I am sort of lost. >> >> I can try again. Begin your sentence, the important utility of this >> construct is demonstrated by the following code: >> >> See if you can complete it. If not, then, then there is no utility. The >> warning says, there is no utility. This isn't a theoretic thing, it is an >> engineering thing. > > Yeah, and as opposed to the non-volatile case removing the volatile > set-but-unused variable even reduces code size! > > Richard.
