------- Comment #10 from ktietz at gcc dot gnu dot org 2009-09-06 20:05 ------- (In reply to comment #9) > (In reply to comment #8) > > (In reply to comment #7) > > > (In reply to comment #6) > > > > I think I found the issue within gfortran for mingw. > > > > > > I think you have this backwards. "the issue within mingw > > > for gfortran" would be a better description of the issue. > > > It is afterall mingw that defines off_t. > > > > > > > Well, in fact it is MS here. But we on mingw-w64 think at the moment about > > to > > add an override option for this by defining _LARGE_FILES. But we have to > > plan > > this in more detail. Would this be an option for you? > > > > If the changes are confined to libgfortran, then I suspect that the > above would be acceptable (although I haven't worked on libgfortran > recently); as long as the changes can be wrapped in '#define __MINGW__' > (or an equivalent). If front-end changes are required, then that would > be much more difficult to accommodate. >
Sorry, I think you misunderstand the suggestion I did. I said that we would allow by defining _LARGE_FILES, that off_t (and ftell, fseek) are defined as POSIX. You have to notice that Windows (and in special mingw targets) aren't POSIX. They are Windows native targets, which have to live with the API MS defines. If the suggested suggestion isn't acceptable, well then files > 2GB won't run for gfortran, sorry. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=40812
