diff options
| author | bors <bors@rust-lang.org> | 2016-11-17 01:18:51 -0800 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2016-11-17 01:18:51 -0800 |
| commit | 29181b3f0c6b27bad24b78630a55b0d527b57068 (patch) | |
| tree | ae83b2ac5692e08195e3fefaa757c8cfc4409d68 /src/rustllvm/RustWrapper.cpp | |
| parent | 6cd5be81cc3a6f972eb00541224037a95a08cc7b (diff) | |
| parent | 48dc6e26ca2ccb72a241ea42af509c4dcf3636fb (diff) | |
| download | rust-29181b3f0c6b27bad24b78630a55b0d527b57068.tar.gz rust-29181b3f0c6b27bad24b78630a55b0d527b57068.zip | |
Auto merge of #37717 - nikomatsakis:region-obligations-pre, r=eddyb
Refactoring towards region obligation
Two refactorings towards the intermediate goal of propagating region obligations through the `InferOk` structure (which in turn leads to the possibility of lazy normalization).
1. Remove `TypeOrigin` and add `ObligationCause`
- as we converge subtyping and obligations and so forth, the ability to keep these types distinct gets harder
2. Propagate obligations from `InferOk` into the surrounding fulfillment context
After these land, I have a separate branch (which still needs a bit of work) that can make the actual change to stop directly adding subregion edges and instead propagate obligations. (This should also make it easier to fix the unsoundness in specialization around lifetimes.)
r? @eddyb
Diffstat (limited to 'src/rustllvm/RustWrapper.cpp')
0 files changed, 0 insertions, 0 deletions