On Wed, Oct 06, 2010 at 08:24:28PM +0400, Anton Korobeynikov wrote:> Jack, > > > This also seems to be the same failure preventing capacita.f90, gas_dyn.f90, induct.f90, protein.f90, rnflow.f90 and > > test_fpu.f90 from compiling. > This might be a long-standing fortran frontend bug where bogus trees > were created. It was specially workarounded in llvm-gcc. >Anton, Have you considered looking up that bug in FSF gcc bugzilla and pinging it? The gfortran developers may have forgotten about it and perhaps it could be addressed by them for gcc 4.6 (as they are still in stage 1 for a few more weeks). Jack> -- > With best regards, Anton Korobeynikov > Faculty of Mathematics and Mechanics, Saint Petersburg State University
Jack,> Have you considered looking up that bug in FSF gcc bugzilla and pinging it? > The gfortran developers may have forgotten about it and perhaps it could be > addressed by them for gcc 4.6 (as they are still in stage 1 for a few more weeks).Well, that bug is in bugzilla since gcc 4.2. Since it does not influence "stock" gcc I doubt they will mark it as high priority. -- With best regards, Anton Korobeynikov Faculty of Mathematics and Mechanics, Saint Petersburg State University
On Wed, Oct 06, 2010 at 09:21:51PM +0400, Anton Korobeynikov wrote:> Jack, > > > Have you considered looking up that bug in FSF gcc bugzilla and pinging it? > > The gfortran developers may have forgotten about it and perhaps it could be > > addressed by them for gcc 4.6 (as they are still in stage 1 for a few more weeks). > Well, that bug is in bugzilla since gcc 4.2. Since it does not > influence "stock" gcc I > doubt they will mark it as high priority.Anton, The gfortran developers have been cleaning up their interactions with the middle end though. Do you have the PR number handy? Jack> > -- > With best regards, Anton Korobeynikov > Faculty of Mathematics and Mechanics, Saint Petersburg State University