------- Comment #2 from dominiq at lps dot ens dot fr 2008-09-02 15:30 ------- > I can believe your comment #5, but I can't believe DW_AT_name is not emitted > at > all (just tried it with x86_64-linux -> i686-darwin9 cross).
I am not sure to understand. If I use egrep with "label.*DW_AT_name" on the assembly file, I get: [ibook-dhum] f90/bug% egrep "label.*DW_AT_name" pr35154-dwarf2.s .ascii "label\0" # DW_AT_name So if "\"label\[^\n\]*\"\[^\n\]*DW_AT_name)" is some Tcl equivalent of "label.*DW_AT_name" for egrep, then it is likely that the test pass. I have tried several variants around (with or without debug/, adding gfortran.dg/) make -k check-gfortran RUNTESTFLAGS="dg.exp=debug/pr35154-dwarf2.f." without being able to test the patched file. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37322