diff options
| author | bors <bors@rust-lang.org> | 2020-02-13 12:53:43 +0000 |
|---|---|---|
| committer | bors <bors@rust-lang.org> | 2020-02-13 12:53:43 +0000 |
| commit | d538b80ad77949e46989cd355cdec193b574f052 (patch) | |
| tree | 2bc1fe9b6da415b4ee049ae12e1a9f43ef4d987d /src/librustc_codegen_llvm/common.rs | |
| parent | be493fe8cc40c3d3f6030a1313c1ff747fce770d (diff) | |
| parent | 33ba83c3a0905c86a6b5182b3b3d35f2f6086e5b (diff) | |
| download | rust-d538b80ad77949e46989cd355cdec193b574f052.tar.gz rust-d538b80ad77949e46989cd355cdec193b574f052.zip | |
Auto merge of #68969 - RalfJung:dont-panic, r=oli-obk
remove Panic variant from InterpError The interpreter engine itself does not raise `Panic` errors any more, so remove them from the error enum. Instead, const-prop and const-eval have to do their own handling of panics. I used the opportunity to refactor the const-eval error handling a bit to use the `MachineStop` variant. Also, in const-prop I could do some cleanup as now, no more lints are being reported in `use_ecx`. However, I am quite puzzled by why exactly the linting there works the way it does -- the code can certainly be cleaned up more, but I don't know enough of the intent to do that. I left some questions for the most confusing parts, but for now behavior should be unchanged by this PR (so, all that weirdness I am asking about is pre-existing and merely maintained here). Cc @wesleywiser Fixes https://github.com/rust-lang/rust/issues/66902 r? @oli-obk
Diffstat (limited to 'src/librustc_codegen_llvm/common.rs')
0 files changed, 0 insertions, 0 deletions
