Re: Fix two ICEs with C++ destructors and LTO

2019-03-19 Thread Rainer Orth
Hi Jan, > these two PRs are about C++ destructors that are not virtual but have > virtual alias. The devirtualization machinery walks from alias to symbol > and is then confused by not knowing the class symbol belongs to. > > I think it would make more sense to avoid these walks but that require

Re: Fix two ICEs with C++ destructors and LTO

2019-03-19 Thread Richard Biener
On Mon, 18 Mar 2019, Jan Hubicka wrote: > Hi, > these two PRs are about C++ destructors that are not virtual but have > virtual alias. The devirtualization machinery walks from alias to symbol > and is then confused by not knowing the class symbol belongs to. > > I think it would make more sense

Fix two ICEs with C++ destructors and LTO

2019-03-18 Thread Jan Hubicka
Hi, these two PRs are about C++ destructors that are not virtual but have virtual alias. The devirtualization machinery walks from alias to symbol and is then confused by not knowing the class symbol belongs to. I think it would make more sense to avoid these walks but that require more of