https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69368
--- Comment #80 from Thomas Koenig <tkoenig at gcc dot gnu.org> --- (In reply to alalaw01 from comment #79) > Is the concern that we can't hide this behind an option, as that would > "drive people away from gfortran" ? If that's the case, can we hide it > behind an option that defaults to pessimization (?? at least for fortran)?? The concern would be if we did not provide a reasonable way around this. I am not in favor of pessimizing valid code by default :-) My preference would be to hide it behind an option, include this option in -std=legacy, and warn from the gfortran front end (with -Wall) if the construct in question (or questionable construct) is encountered. The warning could tell the user which option to select in this case. A one-element array trailing a common block is legal, but its normal use should be quite rare, because it can be replaced without problems with a scalar variable, so I doubt we will generate many false positives. And even if we do, we now have the nice feature of displaying the option which issued the warning, so people should have no problem turning it off.
