Dear All, OK'd by Steve and committed as revision 246256.
Cheers Paul On 27 February 2017 at 19:23, Paul Richard Thomas <paul.richard.tho...@gmail.com> wrote: > Dear All, > > This bug resulted from a cock-up on my part. The mechanism for > suppressing .smod files depended on detecting the presence of a module > procedure by resetting a flag if the module_procedure attribute was > written. Of course, this didn't happen if the module procedure is > private, which rather defeats the requirement that private symbols be > present in an .smod file. The fix traverses the namespace and resets > the flag on finding a module procedure. > > Bootstraps and regtests on FC25/x86_64 - OK for trunk and 6-branch? > > Paul > > 2017-02-27 Paul Thomas <pa...@gcc.gnu.org> > > PR fortran/79676 > * module.c (mio_symbol_attribute): Remove reset of the flag > 'no_module_procedures'. > (check_for_module_procedures): New function. Move declaration > of 'no_module_procedures' to above it. > (gfc_dump_module): Traverse namespace calling new function. > > 2017-02-27 Paul Thomas <pa...@gcc.gnu.org> > > PR fortran/79676 > * gfortran.dg/submodule_28.f08 : New test. -- "If you can't explain it simply, you don't understand it well enough" - Albert Einstein
