Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add documentation on v0 symbol mangling. #97571

Merged
merged 11 commits into from
Jul 28, 2023
Merged

Conversation

ehuss
Copy link
Contributor

@ehuss ehuss commented May 31, 2022

This adds official documentation for the v0 symbol mangling format, migrating the documentation from RFC 2603.
The format was originally stabilized as the -C symbol-mangling-version option, but the specifics were not stabilized (per #90128 (comment)).
Per the discussion at #93661 (comment) this adds those specifics as an official description of the format.

cc #89917

@rust-highfive
Copy link
Collaborator

r? @GuillaumeGomez

(rust-highfive has picked a reviewer for you, use r? to override)

@rust-highfive rust-highfive added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label May 31, 2022
@ehuss
Copy link
Contributor Author

ehuss commented May 31, 2022

Before this is merged, there are a variety of questions to resolve:

  • Is the compiler team ready for this?
  • Are there any open questions about the format?
  • How is it intended to extend this format in the future? For example, if additional const types are stabilized, can v0 just be extended to add them? What constitutes a need for a v1 format? Which parts of v0 can be extended at will?
  • This currently only documents stable parts of the language. Should unstable things like additional const generic types be included now, or just when they are stabilized?
  • The base62 encoding scheme subtly differs from common base62 (or any baseNN) encoding scheme by swapping the uppercase/lowercase letters. Was that intended? ANSWER: Perhaps not, but probably can't be changed now.
  • I can't figure out a way to create a symbol with a placeholder (p) const. How can one be created? ANSWER: Added an example.
  • I left a FIXME in the text for wasm-specific behavior. I'm not sure if that should be included?
  • Should there be a more detailed explanation of the mapping of how Rust entities are encoded? ANSWER: No. This is not designed to be reproducible.

Deviations from the RFC:

  • I defined <bytes> for identifier as UTF-8. Should it be defined as arbitrary bytes instead? rustc_demangle for example does not support non-utf-8. ANSWER: Will leave as UTF-8 for now. If some platform uses non-utf8 symbols, then the format may need to be modified/extended for it.
  • I specified a 64-bit upper-bound for <base-62-number> so that demanglers can have a well-defined limit (and that is how rustc-demangle works). Should I keep that, or should demanglers be prepared for larger values? Removed the limit.
  • I added more detail on the encoding, where the RFC was under-specified. I did my best to make it correct but not over- or under-specified, though it wouldn't hurt for a careful review. For example:
    • fn ABI string converts dashes to underscores.
    • base62 encoding
    • The format for <decimal-number>.
  • I added more detailed recommended demangling (based on rustc-demangle).
  • Adjusted the grammar for decimal-number to make it clear it cannot have leading zeroes.

cc @eddyb @Mark-Simulacrum

@rust-log-analyzer

This comment was marked as resolved.

@eddyb
Copy link
Member

eddyb commented May 31, 2022

Shouldn't this go in the reference? And then have the CLI option link to that?

I don't think we describe language concerns in the rustc CLI docs, and the "Rust v0" mangling is supposed to be the first official Rust language mangling, unlike the rustc-specific "legacy" one.

r? @michaelwoerister cc @lqd

@eddyb
Copy link
Member

eddyb commented May 31, 2022

I'm going to try to do a point-by-point reply to #97571 (comment) (but some of this feels like it's going over things discussed before the RFC was accepted? I suppose those things tend to get lost to time)

  • Is the compiler team ready for this?

    Not speaking for anyone else but I've been operating under the assumption that the RFC itself is the documentation (though it seems reasonable for it to be accessible through the reference, not sure why we didn't just do that), and the main thing we've been waiting on is upstreaming into non-Rust tooling.

  • Are there any open questions about the format?

    Can't think of anything other than the discussion in [RFC2603] Extend <const> to include str and structural constants. rfcs#3161 (which hasn't been merged so I assume it's not part of this PR)

  • How is it intended to extend this format in the future? For example, if additional const types are stabilized, can v0 just be extended to add them? What constitutes a need for a v1 format? Which parts of v0 can be extended at will?

    Generally, as long as leave gaps in the grammar (i.e. unused "letters" in each of 3-4 grammar categories) we can always extend it - by a quick count we have:

    • 7 for path/type (uppercase): GHJLUVW
    • 5 for type (lowercase): gkqrw
    • 23 for "well-known" namespace (uppercase): everything except C (closure), S (shim) and I (impl, but only in a specific polymorphization situation).
    • 21 for implementation-defined namespace (lowercase): this one doesn't really matter, since we just assign them in the compiler and the demangler doesn't have to know about them - worst case we can start putting information into the low bits of the disambiguator if we ever run out

    For constants, [RFC2603] Extend <const> to include str and structural constants. rfcs#3161 both supports the vast majority of types that we could ever have in const generics, and offers (in last paragraph of the PR description) a generalized fallback we could always do in other cases (without updating demanglers at all).

    As for v1, that would require backwards-incompatible changes, i.e. repurposing characters in ways that could accidentally be incorrectly decoded as unrelated v0 manglings. If I had to guess, this would likely a binary format, usable on platforms that do not have the charset limitations, and may even include some kind of compression using a standardized dictionary (zstd was mentioned during the original v0 RFC).

  • This currently only documents stable parts of the language. Should unstable things like additional const generic types be included now, or just when they are stabilized?

    [RFC2603] Extend <const> to include str and structural constants. rfcs#3161 isn't even accepted yet so I don't think there's much of a point in including it now. OTOH, rustc-demangle does implement it, so it might be more of a matter of what we do by default?

  • The base62 encoding scheme subtly differs from common base62 (or any baseNN) encoding scheme by swapping the uppercase/lowercase letters. Was that intended?

    Not sure, I think that was all chosen by @michaelwoerister before I got involved, and I kept it. Though I would've naively assumed lowercase goes before uppercase in any base, so I wouldn't trust myself to have gotten it right either. (Actually, doesn't e.g. base64 look nothing like hex and instead makes A the lowest value digit, so that a bunch of zero bytes turn into AAAAA... in base64? So it's not just the order of lowercase and uppercase)

  • I can't figure out a way to create a symbol with a placeholder (p) const. How can one be created?

    It should be the same as with the type placeholder: impl generics that do not actually apply to the entity being described, like a static, which is never monomorphized by parameters in scope (e.g. mangled / demangled - I...pKpE is ...<_, _>, a type and a const placeholder).

  • I left a FIXME in the text for wasm-specific behavior. I'm not sure if that should be included?

    Maybe, in the general sense of linker symbol documentation, since wasm has a somewhat unique two-level system, but it wouldn't be relevant to a v0 explanation (which IMO should have a clearly marked document of its own, not mixed with other linker symbol concerns like #[no_mangle] and whatnot).

  • Should there be a more detailed explanation of the mapping of how Rust entities are encoded?

    I don't think so, the format isn't designed to be reproducible between two producers, it's more of a "heterogeneous syntax serialization" (honestly we could've probably deduplicated more, or chosen better letters if more time was spent on that front alone, but it's not too bad).

    Deviations from the RFC:

  • I defined <bytes> for identifier as UTF-8. Should it be defined as arbitrary bytes instead? rustc_demangle for example does not support non-utf-8.

    The whole charset is _0-9a-zA-Z (ident-wise, the subset of XID_Start+XID_Continue that overlaps ASCII) so I think <bytes> was just used as a shorthand for that - in theory, it could be UTF-8, but punycode was added specifically so that we could have Unicode w/o UTF-8.

  • I specified a 64-bit upper-bound for <base-62-number> so that demanglers can have a well-defined limit (and that is how rustc-demangle works). Should I keep that, or should demanglers be prepared for larger values?

    Demanglers can have their own implementation limits, I don't think we should put that in the format.
    E.g. rustc-demangle also has like a [char; 128] limit for punycode (because it's #![no_std] so it can't allocate a Vec<char>), and demangling still succeeds, but that overlong identifier is printed out in the standard punycode syntax (used by DNS, primarily).

    For encoding constants I went out of my way to allow "just dump the mangled form" by having it be hex so it just needs a 0x prefix printed (at the cost of compactness compared to base62).

    Honestly, the only reason I can think for disambiguators using base62 is the crate disambiguator hash, and even then, unless it's actually printed, fully decoding it into a number shouldn't be necessary (so rustc-demangle could be updated to make the limit "softer").

  • I added more detail on the encoding, where the RFC was under-specified. I did my best to make it correct but not over- or under-specified, though it wouldn't hurt for a careful review. For example:

    • fn ABI string converts dashes to underscores.
    • base62 encoding
    • The format for <decimal-number>.
  • I added more detailed recommended demangling (based on rustc-demangle).

    For both of these, I'm tempted to say the original RFC should be amended, since more features aren't added, just existing information corrected - there is however some discussion on an existing example of such an amendment so maybe I'm wrong.

@ehuss
Copy link
Contributor Author

ehuss commented Jun 1, 2022

Thank you for the detailed response!

Shouldn't this go in the reference? And then have the CLI option link to that?

Hm, my impression was the opposite, that this is not a part of the language and just an implementation-specific concern. The RFC was accepted by the compiler team, not the lang team. And it isn't really clear what consequence it has for the language if it isn't to be used as a stable ABI or have well-defined translation from Rust entities.

I can imagine in the future that this could potentially provide the basis for a stable ABI, but that seems very far out, no?

v0 explanation (which IMO should have a clearly marked document of its own, not mixed with other linker symbol concerns like #[no_mangle] and whatnot)

Yea, I waffled a bit on how to organize it. I decided to move it to a separate chapter.

Demanglers can have their own implementation limits, I don't think we should put that in the format.

OK, I removed the limit.

Generally, as long as leave gaps in the grammar (i.e. unused "letters" in each of 3-4 grammar categories) we can always extend it

I can understand how something like namespace or basic-type could be extended, but if any of the other open tags were extended, a demangler would have no way to know how to interpret anything that follows. How would that be different from any other backwards-incompatible change? The end result is that an older demangler would be unable to demangle the symbol.

@michaelwoerister
Copy link
Member

Wow, that is some extensive documentation 😀
I'll take a closer look next week.

The base62 encoding scheme subtly differs from common base62 (or any baseNN) encoding scheme by swapping the uppercase/lowercase letters. Was that intended?

Nope, there is no deeper meaning to that. Would have been nice to catch that during RFC review, but I guess now that there are demanglers in external tools, it's too late to change it.

@eddyb
Copy link
Member

eddyb commented Jun 4, 2022

  • Shouldn't this go in the reference? And then have the CLI option link to that?

    Hm, my impression was the opposite, that this is not a part of the language and just an implementation-specific concern. The RFC was accepted by the compiler team, not the lang team. And it isn't really clear what consequence it has for the language if it isn't to be used as a stable ABI or have well-defined translation from Rust entities.

    The exact mapping of source code to symbols is implementation-specific, but any Rust implementation would immediately benefit from generating v0 symbols, as tools exist to demangle them back into Rust syntax, while the format also is able to hide enough redundancy to cover various additional sources of "identity" not explicitly covered otherwise.

    I agree that in retrospect the RFC being strictly compiler team is a bit weird, might be a good idea to poll @rust-lang/lang on how they feel about treating the v0 symbol mangling the way I described ("officially a Rust mangling scheme").

  • I can imagine in the future that this could potentially provide the basis for a stable ABI, but that seems very far out, no?

    Probably only in the vague sense that it's the first sketch we have of being able to serialize what we might consider "definition identity". Definitely agree on far out.

    Earliest similar thing I can think of would be TypeId: use a (v0) mangled type to remain sound in the face of hash collisions. #95845, if we stabilized a method on TypeId to actually get the mangled string (or more likely something that can be decompressed into one), but that PR is not looking favorable right now so that might never happen.

  • Generally, as long as leave gaps in the grammar (i.e. unused "letters" in each of 3-4 grammar categories) we can always extend it

    I can understand how something like namespace or basic-type could be extended, but if any of the other open tags were extended, a demangler would have no way to know how to interpret anything that follows. How would that be different from any other backwards-incompatible change? The end result is that an older demangler would be unable to demangle the symbol.

    Ah I see now. I think there's two levels of incompatibility (not sure how to describe each):

    • symbols using some new feature require new (e.g. type) encoding in mangling
      • only affects code using the new feature, not existing code
      • getting mangling support out there can be part of stabilization
      • we do have the fallback option of choosing a less-appealing-after-demangling encoding, because any syntactic category that can embed paths can represent arbitrary trees
        • e.g. A<B<C, D>, X<Y>> is no different from the sexpr (A (B C D) (X Y)), information-wise
    • new symbols repurpose existing mangling syntax to mean something else
      • this is worse because demanglers can then get confused and "succeed" with the wrong output
      • IOW, this introduces ambiguity, unless some kind of "explicit version bump" is done

@michaelwoerister
Copy link
Member

Great docs, @ehuss! I left some comments in a few places but that's mostly nit-picking.

@eddyb
Copy link
Member

eddyb commented Jun 10, 2022

@michaelwoerister Do any of these comments apply to anything taken from the RFC text? It would be nice to fix them there (since IMO a diff of the RFC and reference should mostly have large-scale reorganization and additions, not e.g. spelling fixes in copied text).

@michaelwoerister
Copy link
Member

@michaelwoerister Do any of these comments apply to anything taken from the RFC text? It would be nice to fix them there (since IMO a diff of the RFC and reference should mostly have large-scale reorganization and additions, not e.g. spelling fixes in copied text).

No, I don't think so.

@JohnCSimon JohnCSimon added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Jul 3, 2022
@JohnCSimon JohnCSimon added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Aug 13, 2022
@JohnCSimon JohnCSimon added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Oct 8, 2022
@rfcbot rfcbot added proposed-final-comment-period Proposed to merge/close by relevant subteam, see T-<team> label. Will enter FCP once signed off. disposition-merge This issue / PR is in PFCP or FCP with a disposition to merge it. labels Jul 3, 2023
@rfcbot rfcbot added final-comment-period In the final comment period and will be merged soon unless new substantive objections are raised. and removed proposed-final-comment-period Proposed to merge/close by relevant subteam, see T-<team> label. Will enter FCP once signed off. labels Jul 13, 2023
@rfcbot
Copy link

rfcbot commented Jul 13, 2023

🔔 This is now entering its final comment period, as per the review above. 🔔

@rfcbot rfcbot added finished-final-comment-period The final comment period is finished for this PR / Issue. and removed final-comment-period In the final comment period and will be merged soon unless new substantive objections are raised. labels Jul 23, 2023
@rfcbot
Copy link

rfcbot commented Jul 23, 2023

The final comment period, with a disposition to merge, as per the review above, is now complete.

As the automated representative of the governance process, I would like to thank the author for their work and everyone else who contributed.

This will be merged soon.

@rfcbot rfcbot added the to-announce Announce this issue on triage meeting label Jul 23, 2023

```text
_RNvNtNtCsgOH4LzxkuMq_7mycrateu8gdel_5qa6escher4bach
││└───┬───┘
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This span extends one character too far.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for spotting that!

@michaelwoerister
Copy link
Member

@ehuss, feel free to r=me when you think it's ready.

@ehuss
Copy link
Contributor Author

ehuss commented Jul 27, 2023

Thanks for the reviews!

@bors r=michaelwoerister rollup

@bors
Copy link
Contributor

bors commented Jul 27, 2023

📌 Commit da4f62e has been approved by michaelwoerister

It is now in the queue for this repository.

@bors bors added S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Jul 27, 2023
workingjubilee added a commit to workingjubilee/rustc that referenced this pull request Jul 27, 2023
…rister

Add documentation on v0 symbol mangling.

This adds official documentation for the v0 symbol mangling format, migrating the documentation from [RFC 2603](https://rust-lang.github.io/rfcs/2603-rust-symbol-name-mangling-v0.html).
The format was originally stabilized as the `-C symbol-mangling-version` option, but the specifics were not stabilized (per rust-lang#90128 (comment)).
Per the discussion at rust-lang#93661 (comment) this adds those specifics as an official description of the format.

cc rust-lang#89917
bors added a commit to rust-lang-ci/rust that referenced this pull request Jul 27, 2023
…kingjubilee

Rollup of 4 pull requests

Successful merges:

 - rust-lang#97571 (Add documentation on v0 symbol mangling.)
 - rust-lang#114122 (tests/ui/hello_world/main.rs: Remove FIXME (rust-lang#62277))
 - rust-lang#114133 (Revert "add tidy check that forbids issue ui test filenames")
 - rust-lang#114139 (Make `--print` with path unstable)

r? `@ghost`
`@rustbot` modify labels: rollup
@workingjubilee
Copy link
Member

workingjubilee commented Jul 28, 2023

I tried to add this to a rollup but somehow it "slipped through" and didn't get actually merged in. I don't know what's up with that but I am assuming that this should be marked
@bors rollup=never

@bors
Copy link
Contributor

bors commented Jul 28, 2023

⌛ Testing commit da4f62e with merge 0ca28c3...

@bors
Copy link
Contributor

bors commented Jul 28, 2023

☀️ Test successful - checks-actions
Approved by: michaelwoerister
Pushing 0ca28c3 to master...

@bors bors added the merged-by-bors This PR was explicitly merged by bors. label Jul 28, 2023
@bors bors merged commit 0ca28c3 into rust-lang:master Jul 28, 2023
@rustbot rustbot added this to the 1.73.0 milestone Jul 28, 2023
@rust-timer
Copy link
Collaborator

Finished benchmarking commit (0ca28c3): comparison URL.

Overall result: no relevant changes - no action needed

@rustbot label: -perf-regression

Instruction count

This benchmark run did not return any relevant results for this metric.

Max RSS (memory usage)

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

mean range count
Regressions ❌
(primary)
0.6% [0.6%, 0.6%] 1
Regressions ❌
(secondary)
2.3% [2.3%, 2.3%] 1
Improvements ✅
(primary)
- - 0
Improvements ✅
(secondary)
-2.8% [-3.5%, -2.0%] 2
All ❌✅ (primary) 0.6% [0.6%, 0.6%] 1

Cycles

This benchmark run did not return any relevant results for this metric.

Binary size

This benchmark run did not return any relevant results for this metric.

Bootstrap: 653.092s -> 653.16s (0.01%)

@apiraino apiraino removed the to-announce Announce this issue on triage meeting label Aug 4, 2023
wip-sync pushed a commit to NetBSD/pkgsrc-wip that referenced this pull request Oct 6, 2023
Language
--------

- [Uplift `clippy::fn_null_check` lint as `useless_ptr_null_checks`.]
  (rust-lang/rust#111717)
- [Make `noop_method_call` warn by default.]
  (rust-lang/rust#111916)
- [Support interpolated block for `try` and `async` in macros.]
  (rust-lang/rust#112953)
- [Make `unconditional_recursion` lint detect recursive drops.]
  (rust-lang/rust#113902)
- [Future compatibility warning for some impls being incorrectly
  considered not overlapping.]
  (rust-lang/rust#114023)
- [The `invalid_reference_casting` lint is now **deny-by-default**
  (instead of allow-by-default)]
  (rust-lang/rust#112431)

Compiler
--------

- [Write version information in a `.comment` section like GCC/Clang.]
  (rust-lang/rust#97550)
- [Add documentation on v0 symbol mangling.]
  (rust-lang/rust#97571)
- [Stabilize `extern "thiscall"` and `"thiscall-unwind"` ABIs.]
  (rust-lang/rust#114562)
- [Only check outlives goals on impl compared to trait.]
  (rust-lang/rust#109356)
- [Infer type in irrefutable slice patterns with fixed length as array.]
  (rust-lang/rust#113199)
- [Discard default auto trait impls if explicit ones exist.]
  (rust-lang/rust#113312)
- Add several new tier 3 targets:
    - [`aarch64-unknown-teeos`]
      (rust-lang/rust#113480)
    - [`csky-unknown-linux-gnuabiv2`]
      (rust-lang/rust#113658)
    - [`riscv64-linux-android`]
      (rust-lang/rust#112858)
    - [`riscv64gc-unknown-hermit`]
      (rust-lang/rust#114004)
    - [`x86_64-unikraft-linux-musl`]
      (rust-lang/rust#113411)
    - [`x86_64-unknown-linux-ohos`]
      (rust-lang/rust#113061)
- [Add `wasm32-wasi-preview1-threads` as a tier 2 target.]
  (rust-lang/rust#112922)

Refer to Rust's [platform support page][platform-support-doc]
for more information on Rust's tiered platform support.

Libraries
---------

- [Add `Read`, `Write` and `Seek` impls for `Arc<File>`.]
  (rust-lang/rust#94748)
- [Merge functionality of `io::Sink` into `io::Empty`.]
  (rust-lang/rust#98154)
- [Implement `RefUnwindSafe` for `Backtrace`]
  (rust-lang/rust#100455)
- [Make `ExitStatus` implement `Default`]
  (rust-lang/rust#106425)
- [`impl SliceIndex<str> for (Bound<usize>, Bound<usize>)`]
  (rust-lang/rust#111081)
- [Change default panic handler message format.]
  (rust-lang/rust#112849)
- [Cleaner `assert_eq!` & `assert_ne!` panic messages.]
  (rust-lang/rust#111071)
- [Correct the (deprecated) Android `stat` struct definitions.]
  (rust-lang/rust#113130)

Stabilized APIs
---------------

- [Unsigned `{integer}::div_ceil`]
  (https://doc.rust-lang.org/stable/std/primitive.u32.html#method.div_ceil)
- [Unsigned `{integer}::next_multiple_of`]
  (https://doc.rust-lang.org/stable/std/primitive.u32.html#method.next_multiple_of)
- [Unsigned `{integer}::checked_next_multiple_of`]
  (https://doc.rust-lang.org/stable/std/primitive.u32.html#method.checked_next_multiple_of)
- [`std::ffi::FromBytesUntilNulError`]
  (https://doc.rust-lang.org/stable/std/ffi/struct.FromBytesUntilNulError.html)
- [`std::os::unix::fs::chown`]
  (https://doc.rust-lang.org/stable/std/os/unix/fs/fn.chown.html)
- [`std::os::unix::fs::fchown`]
  (https://doc.rust-lang.org/stable/std/os/unix/fs/fn.fchown.html)
- [`std::os::unix::fs::lfchown`]
  (https://doc.rust-lang.org/stable/std/os/unix/fs/fn.lchown.html)
- [`LocalKey::<Cell<T>>::get`]
  (https://doc.rust-lang.org/stable/std/thread/struct.LocalKey.html#method.get)
- [`LocalKey::<Cell<T>>::set`]
  (https://doc.rust-lang.org/stable/std/thread/struct.LocalKey.html#method.set)
- [`LocalKey::<Cell<T>>::take`]
  (https://doc.rust-lang.org/stable/std/thread/struct.LocalKey.html#method.take)
- [`LocalKey::<Cell<T>>::replace`]
  (https://doc.rust-lang.org/stable/std/thread/struct.LocalKey.html#method.replace)
- [`LocalKey::<RefCell<T>>::with_borrow`]
  (https://doc.rust-lang.org/stable/std/thread/struct.LocalKey.html#method.with_borrow)
- [`LocalKey::<RefCell<T>>::with_borrow_mut`]
  (https://doc.rust-lang.org/stable/std/thread/struct.LocalKey.html#method.with_borrow_mut)
- [`LocalKey::<RefCell<T>>::set`]
  (https://doc.rust-lang.org/stable/std/thread/struct.LocalKey.html#method.set-1)
- [`LocalKey::<RefCell<T>>::take`]
  (https://doc.rust-lang.org/stable/std/thread/struct.LocalKey.html#method.take-1)
- [`LocalKey::<RefCell<T>>::replace`]
  (https://doc.rust-lang.org/stable/std/thread/struct.LocalKey.html#method.replace-1)

These APIs are now stable in const contexts:

- [`rc::Weak::new`]
  (https://doc.rust-lang.org/stable/alloc/rc/struct.Weak.html#method.new)
- [`sync::Weak::new`]
  (https://doc.rust-lang.org/stable/alloc/sync/struct.Weak.html#method.new)
- [`NonNull::as_ref`]
  (https://doc.rust-lang.org/stable/core/ptr/struct.NonNull.html#method.as_ref)

Cargo
-----

- [Encode URL params correctly for `SourceId` in `Cargo.lock`.]
  (rust-lang/cargo#12280)
- [Bail out an error when using `cargo::` in custom build script.]
  (rust-lang/cargo#12332)

Compatibility Notes
-------------------

- [Update the minimum external LLVM to 15.]
  (rust-lang/rust#114148)
- [Check for non-defining uses of return position `impl Trait`.]
  (rust-lang/rust#112842)

Internal Changes
----------------

These changes do not affect any public interfaces of Rust, but they represent
significant improvements to the performance or internals of rustc and related
tools.

- [Remove LLVM pointee types, supporting only opaque pointers.]
  (rust-lang/rust#105545)
- [Port PGO/LTO/BOLT optimized build pipeline to Rust.]
  (rust-lang/rust#112235)
- [Replace in-tree `rustc_apfloat` with the new version of the crate.]
  (rust-lang/rust#113843)
- [Update to LLVM 17.]
  (rust-lang/rust#114048)
- [Add `internal_features` lint for internal unstable features.]
  (rust-lang/rust#108955)
- [Mention style for new syntax in tracking issue template.]
  (rust-lang/rust#113586)
netbsd-srcmastr pushed a commit to NetBSD/pkgsrc that referenced this pull request Nov 16, 2023
Pkgsrc changes:
 * Adjust patches and cargo checksums to new versions.
 * For an external LLVM, set dependency of llvm >= 15, in accordance
   with the upstream changes.
 * Add a patch with a backport from LLVM 17.0.3 fixing codegen for
   PPC, ref. rust-lang/rust#116845

Upstream changes:

Version 1.73.0 (2023-10-05)
==========================

Language
--------

- [Uplift `clippy::fn_null_check` lint as `useless_ptr_null_checks`.]
  (rust-lang/rust#111717)
- [Make `noop_method_call` warn by default.]
  (rust-lang/rust#111916)
- [Support interpolated block for `try` and `async` in macros.]
  (rust-lang/rust#112953)
- [Make `unconditional_recursion` lint detect recursive drops.]
  (rust-lang/rust#113902)
- [Future compatibility warning for some impls being incorrectly
  considered not overlapping.]
  (rust-lang/rust#114023)
- [The `invalid_reference_casting` lint is now **deny-by-default**
  (instead of allow-by-default)]
  (rust-lang/rust#112431

Compiler
--------

- [Write version information in a `.comment` section like GCC/Clang.]
  (rust-lang/rust#97550)
- [Add documentation on v0 symbol mangling.]
  (rust-lang/rust#97571)
- [Stabilize `extern "thiscall"` and `"thiscall-unwind"` ABIs.]
  (rust-lang/rust#114562)
- [Only check outlives goals on impl compared to trait.]
  (rust-lang/rust#109356)
- [Infer type in irrefutable slice patterns with fixed length as array.]
  (rust-lang/rust#113199)
- [Discard default auto trait impls if explicit ones exist.]
  (rust-lang/rust#113312)
- Add several new tier 3 targets:
    - [`aarch64-unknown-teeos`]
      (rust-lang/rust#113480)
    - [`csky-unknown-linux-gnuabiv2`]
      (rust-lang/rust#113658)
    - [`riscv64-linux-android`]
      (rust-lang/rust#112858)
    - [`riscv64gc-unknown-hermit`]
      (rust-lang/rust#114004)
    - [`x86_64-unikraft-linux-musl`]
      (rust-lang/rust#113411)
    - [`x86_64-unknown-linux-ohos`]
      (rust-lang/rust#113061)
- [Add `wasm32-wasi-preview1-threads` as a tier 2 target.]
  (rust-lang/rust#112922)

Refer to Rust's [platform support page][platform-support-doc]
for more information on Rust's tiered platform support.

Libraries
---------

- [Add `Read`, `Write` and `Seek` impls for `Arc<File>`.]
  (rust-lang/rust#94748)
- [Merge functionality of `io::Sink` into `io::Empty`.]
  (rust-lang/rust#98154)
- [Implement `RefUnwindSafe` for `Backtrace`]
  (rust-lang/rust#100455)
- [Make `ExitStatus` implement `Default`]
  (rust-lang/rust#106425)
- [`impl SliceIndex<str> for (Bound<usize>, Bound<usize>)`]
  (rust-lang/rust#111081)
- [Change default panic handler message format.]
  (rust-lang/rust#112849)
- [Cleaner `assert_eq!` & `assert_ne!` panic messages.]
  (rust-lang/rust#111071)
- [Correct the (deprecated) Android `stat` struct definitions.]
  (rust-lang/rust#113130)

Stabilized APIs
---------------

- [Unsigned `{integer}::div_ceil`]
  (https://doc.rust-lang.org/stable/std/primitiv e.u32.html#method.div_ceil)
- [Unsigned `{integer}::next_multiple_of`]
  (https://doc.rust-lang.org/stable/std/primitive.u32.html#method.next_multiple_of)
- [Unsigned `{integer}::checked_next_multiple_of`]
  (https://doc.rust-lang.org/stable/std/primitive.u32.html#method.checked_next_multiple_of)
- [`std::ffi::FromBytesUntilNulError`]
  (https://doc.rust-lang.org/stable/std/ffi/struct.FromBytesUntilNulError.html)
- [`std::os::unix::fs::chown`]
  (https://doc.rust-lang.org/stable/std/os/unix/fs/fn.chown.html)
- [`std::os::unix::fs::fchown`]
  (https://doc.rust-lang.org/stable/std/os/unix/fs/fn.fchown.html)
- [`std::os::unix::fs::lfchown`]
  (https://doc.rust-lang.org/stable/std/os/unix/fs/fn.lchown.html)
- [`LocalKey::<Cell<T>>::get`]
  (https://doc.rust-lang.org/stable/std/thread/struct.LocalKey.html#method.get)
- [`LocalKey::<Cell<T>>::set`]
  (https://doc.rust-lang.org/stable/std/thread/struct.LocalKey.html#method.set)
- [`LocalKey::<Cell<T>>::take`]
  (https://doc.rust-lang.org/stable/std/thread/struct.LocalKey.html#method.take)
- [`LocalKey::<Cell<T>>::replace`]
  (https://doc.rust-lang.org/stable/std/thread/struct.LocalKey.html#method.replace)
- [`LocalKey::<RefCell<T>>::with_borrow`]
  (https://doc.rust-lang.org/stable/std/thread/struct.LocalKey.html#method.with_borrow)
- [`LocalKey::<RefCell<T>>::with_borrow_mut`]
  (https://doc.rust-lang.org/stable/std/thread/struct.LocalKey.html#method.with_borrow_mut)
- [`LocalKey::<RefCell<T>>::set`]
  (https://doc.rust-lang.org/stable/std/thread/struct.LocalKey.html#method.set-1)
- [`LocalKey::<RefCell<T>>::take`]
  (https://doc.rust-lang.org/stable/std/thread/struct.LocalKey.html#method.take-1)
- [`LocalKey::<RefCell<T>>::replace`]
  (https://doc.rust-lang.org/stable/std/thread/struct.LocalKey.html#method.replace-1)

These APIs are now stable in const contexts:

- [`rc::Weak::new`]
  (https://doc.rust-lang.org/stable/alloc/rc/struct.Weak.html#method.new)
- [`sync::Weak::new`]
  (https://doc.rust-lang.org/stable/alloc/sync/struct.Weak.html#method.new)
- [`NonNull::as_ref`]
  (https://doc.rust-lang.org/stable/core/ptr/struct.NonNull.html#method.as_ref)

Cargo
-----

- [Encode URL params correctly for `SourceId` in `Cargo.lock`.]
  (rust-lang/cargo#12280)
- [Bail out an error when using `cargo::` in custom build script.]
  (rust-lang/cargo#12332)

Misc
----

Compatibility Notes
-------------------

- [Update the minimum external LLVM to 15.]
  (rust-lang/rust#114148)
- [Check for non-defining uses of return position `impl Trait`.]
  (rust-lang/rust#112842)

Internal Changes
----------------

These changes do not affect any public interfaces of Rust, but they
represent significant improvements to the performance or internals
of rustc and related tools.

- [Remove LLVM pointee types, supporting only opaque pointers.]
  (rust-lang/rust#105545)
- [Port PGO/LTO/BOLT optimized build pipeline to Rust.]
  (rust-lang/rust#112235)
- [Replace in-tree `rustc_apfloat` with the new version of the crate.]
  (rust-lang/rust#113843)
- [Update to LLVM 17.]
  (rust-lang/rust#114048)
- [Add `internal_features` lint for internal unstable features.]
  (rust-lang/rust#108955)
- [Mention style for new syntax in tracking issue template.]
  (rust-lang/rust#113586)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
disposition-merge This issue / PR is in PFCP or FCP with a disposition to merge it. finished-final-comment-period The final comment period is finished for this PR / Issue. merged-by-bors This PR was explicitly merged by bors. S-waiting-on-bors Status: Waiting on bors to run and complete tests. Bors will change the label on completion. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue.
Projects
None yet
Development

Successfully merging this pull request may close these issues.