Disable field reordering for repr(int). #56887
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
(rust_highfive has picked a reviewer for you, use r? to override) |
rust-highfive
added
the
S-waiting-on-review
|
r? @eddyb |
|
The order of commits is the opposite github shows because github doesn't show the topological order... :( Anyway. cc @BorisChiou, I think this is the problem you hit with cbindgen that time. |
This comment has been minimized.
This comment has been minimized.
RFC rust-lang#2195 specifies that a repr(int) enum such as: #[repr(u8)] enum MyEnum { B { x: u8, y: i16, z: u8 }, } has a layout that is equivalent to: #[repr(C)] enum MyEnumVariantB { tag: u8, x: u8, y: i16, z: u8 }, However this isn't actually implemented, with the actual layout being roughly equivalent to: union MyEnumPayload { B { x: u8, y: i16, z: u8 }, } #[repr(packed)] struct MyEnum { tag: u8, payload: MyEnumPayload, } Thus the variant payload is *not* subject to repr(C) ordering rules, and gets re-ordered as `{ x: u8, z: u8, z: i16 }` The existing tests added in pull-req rust-lang#45688 fail to catch this as the repr(C) ordering just happens to match the current Rust ordering in this case; adding a third field reveals the problem.
emilio
force-pushed
the
enum-field-reordering
branch
from
c5dbc0a
to
d84bdba
Compare
eddyb
reviewed
Dec 19, 2018
| pub fn inhibit_struct_field_reordering_opt(&self) -> bool { | ||
| !(self.flags & ReprFlags::IS_UNOPTIMISABLE).is_empty() || (self.pack == 1) | ||
| self.flags.intersects(ReprFlags::IS_UNOPTIMISABLE) || self.pack == 1 || | ||
| self.int.is_some() |
There was a problem hiding this comment.
cc @rkruppe @nagisa This seems somewhat fragile (I mean the pre-existing code).
The check for pack == 1 means things will still be reordered with #[repr(packed(N))], is that intentional?
Also, it seems to me that inhibit_enum_layout_opt is a subset of this function and maybe we should have just one function that checks all conditions?
There was a problem hiding this comment.
The repr RFC does not specify any behaviour /wrt repr(packed(N)) for enums and I’m not sure I recall us guaranteeing that repr(packed) specifies any representation and the reference seems to agree.
There was a problem hiding this comment.
In fact as per strict reading of the reference, self.pack == 1 inhibiting reordering would be somewhat wrong, as it should cause "packing" for the original ordering, but since the "default" ordering is not specified, it seems to be fine either way...
|
@bors r+ |
|
📌 Commit d84bdba has been approved by |
bors
added
S-waiting-on-bors
pietroalbini
added a commit
to pietroalbini/rust
that referenced
this pull request
Dec 20, 2018
Disable field reordering for repr(int). This fixes the problem that the test in rust-lang#56619 uncovers. Closes rust-lang#56619.
Mark-Simulacrum
added a commit
to Mark-Simulacrum/rust
that referenced
this pull request
Dec 21, 2018
Disable field reordering for repr(int). This fixes the problem that the test in rust-lang#56619 uncovers. Closes rust-lang#56619.
pietroalbini
added a commit
to pietroalbini/rust
that referenced
this pull request
Dec 21, 2018
Disable field reordering for repr(int). This fixes the problem that the test in rust-lang#56619 uncovers. Closes rust-lang#56619.
bors
added a commit
that referenced
this pull request
Dec 21, 2018
Rollup of 20 pull requests Successful merges: - #56802 (Add DoubleEndedIterator::nth_back) - #56842 (Add unstable VecDeque::rotate_{left|right}) - #56869 (Reduce search-index.js size) - #56887 (Disable field reordering for repr(int).) - #56892 (rustc: Update Clang used to build LLVM on Linux) - #56909 (static eval: Do not ICE on layout size overflow) - #56914 (Ignore ui/target-feature-gate on sparc, sparc64, powerpc, powerpc64 and powerpc64le) - #56917 (Simplify MIR generation for logical operations) - #56919 (Remove a wrong multiplier on relocation offset computation) - #56933 (Add --progress to git submodule commands in x.py) - #56941 (deny intra-doc link resolution failures in libstd) - #56964 (Remove `TokenStream::JointTree`.) - #56970 (Mem uninit doc ptr drop) - #56973 (make basic CTFE tracing available on release builds) - #56979 (Adding unwinding support for x86_64_fortanix_unknown_sgx target.) - #56981 (miri: allocation is infallible) - #56984 (A few tweaks to dropck_outlives) - #56989 (Fix compiletest `trim` deprecation warnings) - #56992 (suggest similar lint names for unknown lints) - #57002 (Stabilize Vec(Deque)::resize_with) Failed merges: r? @ghost
bors
added a commit
that referenced
this pull request
Dec 21, 2018
Rollup of 20 pull requests Successful merges: - #56802 (Add DoubleEndedIterator::nth_back) - #56842 (Add unstable VecDeque::rotate_{left|right}) - #56869 (Reduce search-index.js size) - #56887 (Disable field reordering for repr(int).) - #56892 (rustc: Update Clang used to build LLVM on Linux) - #56909 (static eval: Do not ICE on layout size overflow) - #56914 (Ignore ui/target-feature-gate on sparc, sparc64, powerpc, powerpc64 and powerpc64le) - #56917 (Simplify MIR generation for logical operations) - #56919 (Remove a wrong multiplier on relocation offset computation) - #56933 (Add --progress to git submodule commands in x.py) - #56941 (deny intra-doc link resolution failures in libstd) - #56964 (Remove `TokenStream::JointTree`.) - #56970 (Mem uninit doc ptr drop) - #56973 (make basic CTFE tracing available on release builds) - #56979 (Adding unwinding support for x86_64_fortanix_unknown_sgx target.) - #56981 (miri: allocation is infallible) - #56984 (A few tweaks to dropck_outlives) - #56989 (Fix compiletest `trim` deprecation warnings) - #56992 (suggest similar lint names for unknown lints) - #57002 (Stabilize Vec(Deque)::resize_with) Failed merges: r? @ghost
Centril
added a commit
to Centril/rust
that referenced
this pull request
Dec 22, 2018
Disable field reordering for repr(int). This fixes the problem that the test in rust-lang#56619 uncovers. Closes rust-lang#56619.
|
@bors p=30 (rollup fairness) |
|
☀️ Test successful - status-appveyor, status-travis |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This fixes the problem that the test in #56619 uncovers.
Closes #56619.