expose drop_in_place as ptr::drop_in_place #27204
Conversation
|
r? @pcwalton (rust_highfive has picked a reviewer for you, use r? to override) |
| /// # Undefined Behaviour | ||
| /// | ||
| /// This has all the same safety problems as `ptr::read` with respect to | ||
| /// invalid pointers, types, and double drops. |
There was a problem hiding this comment.
extra tabbing in these comments
alexcrichton
added
I-needs-decision
|
The outstanding issue for this is ptr::drop_in_place vs mem::drop_in_place and whether it should take a *mut or an &mut. mem seems more consistent with mem::drop, but ptr is strictly more flexible (&mut coerces to *mut). |
|
(This will need a tracking issue and an |
| @@ -39,6 +39,9 @@ pub use intrinsics::copy; | |||
| #[stable(feature = "rust1", since = "1.0.0")] | |||
| pub use intrinsics::write_bytes; | |||
|
|
|||
| #[unstable(feature = "drop_in_place", reason = "just exposed, needs approval")] | |||
There was a problem hiding this comment.
In addition to what @huonw said about this needing issue= now this'll also need to be applied to the definition in the intrinsics module itself to work (we don't look at reexports themselves)
|
r=me with @huonw's and my tweaks to the stability attribute |
Gankra
force-pushed
the
drop-in
branch
2 times, most recently
from
afbe8ec to
46986bf
Compare
|
@bors r=alexcrichton |
|
📌 Commit 46986bf has been approved by |
|
@bors r=alexcrichton (making the feature gate actually do a thing broke stuff) |
|
📌 Commit be29f00 has been approved by |
|
fsk how did that conflict even happen |
|
@bors r- |
|
@bors r=alexcrichton (This branch was super old, that's how!) |
|
📌 Commit 5c6f2c5 has been approved by |
|
(looks like travis is unhappy) |
|
augh run-pass test using the intrinsic |
|
@bors r- |
|
@bors r=alexcrichton |
|
📌 Commit 3547e8c has been approved by |
|
(actually did a local make-check) |
|
⌛ Testing commit 3547e8c with merge 58d247a... |
|
💔 Test failed - auto-win-msvc-64-opt |
|
☔ The latest upstream changes (presumably #28306) made this pull request unmergeable. Please resolve the merge conflicts. |
|
What's the status of this? |
|
@apasel422 I'm kinda AWOL; feel free to rebase if you want this to land! |
|
Will do! |
This is a rebase of #27204. r? @alexcrichton CC @gankro
This intrinsic is necessary to drop DSTs and should generally be the way to drop memory before manually deallocating it. As such, it should be properly exposed by the standard library.