librustc: emit loop for expr_repeat instead of 2n instructions in [x, n] #5524
Conversation
|
Sorry for all the criticism, but you did steal what I was trying to do. ;) (Though you did it far better than me.) I've got a branch with the changes I suggested: https://github.com/dbaupp/rust/commit/6d50851bd005a942a70a052c4cd7327d8ab4b749. (Also, something I struggled with: do you know if it is safe to drop the cleanups? I have no idea what their purpose is.) |
|
@dbaupp Haha, thanks for the suggestions. I just applied the commit from your branch. Regarding the cleanups, it seems to work fine as it but I did try a few things which didn't work out. It'd be good if someone more knowledge on the subject could comment. (@pcwalton ?) |
|
I'd like to see a test that tests something like Regarding whether the temp cleanups are needed, I think they're needed in the case in which failure happens during a copy. I think that, due to the fact that copy constructors are not overridable in Rust, this can only happen due to OOM (which probably causes all sorts of bad things to happen in Rust anyway). So I think it's probably OK to omit them, but it should be noted with a comment. |
`let v = [24, ..1000];` now more or less emits the same IR as:
```Rust
let mut i = 0;
while i < 1000 {
v[i] = 24;
i += 1;
}
```
LLVM will still turn it into a memset if possible with optimization on.
map_clone: avoid suggesting `copied()` for &mut changelog: map_clone: avoid suggesting `copied()` for &mut Fixes rust-lang#5524
let v = [24, ..1000];now more or less emits the same IR as:LLVM will still turn it into a memset if possible with optimization on.