------- Additional Comments From mmitchel at gcc dot gnu dot org 2005-09-27 16:32 ------- I don't see what the alleged front-end bug is in this PR.
The layout I would expect for C on IA32 GNU/Linux would be: 0: B's vptr 4: i0 8: i1 12: c 16: A's vptr 20: i 24: c0 25: c1 That is exactly what I seem to see when looking at the output of the class layout algorithm. In the dump in comment #12 I see no indication of overlap between "_vptr.A" (which is located at offset 128, as predicated above) and "i0" (which is, strangely, listed twice at offset zero, which is not correct). I see no evidence in looking at the TYPE_FIELDS list that i0 is at offset zero. Please explain exactly what you think the front end defect is here, in terms of front end data structures. What FIELD_DECL has an incorrect offset? -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=22488
