https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114541
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Ever confirmed|0 |1
Status|UNCONFIRMED |NEW
Last reconfirmed| |2024-12-09
Keywords|accepts-invalid |
--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
So we're now no longer "accepting" the invalid code but we're still ICEing. As
said during the review we should think of how to handle labels that are nowhere
defined for !CFG and how to handle computed goto, which even with CFG has sth
different than __BBN. Only after that we should see how to handle error
recovery better.
Error recovery is IMO most important when parsing the output of
-fdump-tree-*-gimple, but not for "random" invalid input.
We now ICE like the following because of missing edges.
t.c: In function ‘foo’:
t.c:24:1: internal compiler error: in calc_dfs_tree, at dominance.cc:458
24 | }
| ^
0x37ca710 internal_error(char const*, ...)
/space/rguenther/src/gcc/gcc/diagnostic-global-context.cc:517
0x379a1d6 fancy_abort(char const*, int, char const*)
/space/rguenther/src/gcc/gcc/diagnostic.cc:1704
0x11be960 calc_dfs_tree
/space/rguenther/src/gcc/gcc/dominance.cc:458
0x11bf5c6 calculate_dominance_info(cdi_direction, bool)
/space/rguenther/src/gcc/gcc/dominance.cc:743
0xf82f88 c_parser_parse_gimple_body(c_parser*, char*, c_declspec_il,
profile_count)
/space/rguenther/src/gcc/gcc/c/gimple-parser.cc:321
