This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Fix PR debug/14114
On Mar 3, 2004, at 2:12 PM, Zack Weinberg wrote:
Eric Botcazou <ebotcazou@libertysurf.fr> writes:
I don't see why this happens - you are saying that it is the *same*
tree that was already sent from the compilation of 'c', so the
front-end should recognize that it doesn't have to be done again.
Why? Is there a special flag that is set on a definition DECL which
has
already been dwarfized?
I don't know, honestly, but consider that the same problem does not
occur for
void c();
void c() { ... }
so I think the real bug will be found in where this case is treated
differently from your case.
gen_subprogram_die() in dwarf2out.c is prepared to handle this case.
When same tree is supplied again, it checks that previous die generated
for the tree was declaration die. If possible it reuses that declaration
die for definition.
gen_subprogram_die() is not prepared to handle the sequence presented
by pr14114, that is definition first and then declaration.
--
Devang