------- Comment #54 from rguenther at suse dot de 2007-10-27 20:03 ------- Subject: Re: wrong types in character array/scalar binop
On Sat, 27 Oct 2007, burnus at gcc dot gnu dot org wrote: > ------- Comment #53 from burnus at gcc dot gnu dot org 2007-10-27 19:57 > ------- > > Note that I still see achar_4.f90 fail with type-checking and there are now > > some more testcases that also fail. > > Reopened based on this comment to make sure we won't forget about this PR. > > To recap: > > a) We need to fix the test "gfortran.dg/char_cast_1.f90" to allow for > "S$5", "__S_5", "S___5" besides "S.5". > > b) We need to fix get the types right. Richard, can you pin-point where the > types are still wrong? > How did you check for the miss-matched types? If you run the testsuite for a stage1 compiler (or if you enable type-checking with --enable-checking=yes,types) you'll get gfortran failures (ICE due to wrong types). For example /space/rguenther/src/svn/pointer_plus/gcc/testsuite/gfortran.dg/achar_4.f90:8: internal compiler error: verify_gimple failed^M Please submit a full bug report,^M with preprocessed source if appropriate.^M See <http://gcc.gnu.org/bugs.html> for instructions.^M compiler exited with status 1 output is: /space/rguenther/src/svn/pointer_plus/gcc/testsuite/gfortran.dg/achar_4.f90: In function 'up':^M /space/rguenther/src/svn/pointer_plus/gcc/testsuite/gfortran.dg/achar_4.f90:8: error: non-trivial conversion at assignment^M char[1:1]^M char^M (*D.1000)[S.24] = D.1018^M Richard. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31608