Attempting to regenerate gcc.pot failed with: /usr/bin/xgettext: Non-ASCII string at fortran/resolve.c:9894. Please specify the source encoding through --from-code.
The code was using a UTF-8 ligature for "fi", I suppose resulting from a cut-and-paste from some typedef document using that ligature. I've committed as obvious this patch to fix it to use normal ASCII letters instead. Index: fortran/ChangeLog =================================================================== --- fortran/ChangeLog (revision 196241) +++ fortran/ChangeLog (working copy) @@ -1,3 +1,8 @@ +2013-02-24 Joseph Myers <jos...@codesourcery.com> + + * resolve.c (generate_component_assignments): Don't use UTF-8 + ligature in diagnostic. + 2013-02-21 Janus Weil <ja...@gcc.gnu.org> PR fortran/56385 Index: fortran/resolve.c =================================================================== --- fortran/resolve.c (revision 196241) +++ fortran/resolve.c (working copy) @@ -9891,7 +9891,7 @@ (*code)->expr1->rank ? 1 : 0); if (depth > 1) { - gfc_warning ("TODO: type-bound de???ned assignment(s) at %L not " + gfc_warning ("TODO: type-bound defined assignment(s) at %L not " "done because multiple part array references would " "occur in intermediate expressions.", &(*code)->loc); return; -- Joseph S. Myers jos...@codesourcery.com