Properly allow macro expanded `format_args` invocations to uses captures #106505

Nilstrieb · 2023-01-05T19:29:58Z

Originally, this was kinda half-allowed. There were some primitive checks in place that looked at the span to see whether the input was likely a literal. These "source literal" checks are needed because the spans created during format_args parsing only make sense when it is indeed a literal that was written in the source code directly.

This is orthogonal to the restriction that the first argument must be a "direct literal", not being exanpanded from macros. This restriction was imposed by RFC 2795 on the basis of being too confusing. But this was only concerned with the argument of the invocation being a literal, not whether it was a source literal (maybe in spirit it meant it being a source literal, this is not clear to me).

Since the original check only really cared about source literals (which is good enough to deny the format_args!(concat!()) example), macros expanding to format_args invocations were able to use implicit captures if they spanned the string in a way that lead back to a source string.

The "source literal" checks were not strict enough and caused ICEs in certain cases (see #106191). So I tightened it up in #106195 to really only work if it's a direct source literal.

This caused the indoc crate to break. indoc transformed the source literal by removing whitespace, which made it not a "source literal" anymore (which is required to fix the ICE). But since indoc spanned the literal in ways that made the old check think that it's a literal, it was able to use implicit captures (which is useful and nice for the users of indoc).

This commit properly seperates the previously introduced concepts of "source literal" and "direct literal" and therefore allows indoc invocations, which don't create "source literals" to use implicit captures again.

Fixes #106191

Nilstrieb · 2023-01-05T19:38:37Z

Since this changes what kind of code is allowed to compile and settles a question not fully answered by the RFC, I'll lang nominate this.

The PR description isn't really relevant here, TLDR: The check for "the first argument must be a string literal" was implemented badly and a recent PR of mine made it worse, forcing me to properly fix it here.

Question for the lang team

Should format_args! invocations that were generated by a proc-macro be able to use implicit argument captures?

This used to be kind of half-supported on stable (depending on the spans of the string literal) and is now (accidentally) mostly unsupported on nightly.

A concrete example that does not compile before this PR (neither on stable nor nightly) but does compile with it:
pm.rs

extern crate proc_macro;

#[proc_macro]
pub fn foo(_: proc_macro::TokenStream) -> proc_macro::TokenStream {
    r#"{ let x = 5; format!("{x}") }"#.parse().unwrap()
}

rustc pm.rs --crate-type proc-macro

fn main() {
    dbg!(pm::foo!());
}

rustc main.rs --extern pm=libpm.so

Arguments for it

The indoc crate is an example of a crate that makes usage of this. Note that format args captures indoc compiles on stable but not nightly because of the old broken and inconsistent semantics. It's in my opinion unreasonable to support the indoc usage but not my example above (since only the spans of the string literal are different).

Arguments against it

This could lead to confusing behavior in contrived situations. indoc (or other crates like it) could do their own parsing of the implicit captures and expand them out into separate bindings before forwarding them to format_args, but this can of course be said for the whole format_args machinery.

tmandry · 2023-01-11T21:32:47Z

Discussed in yesterday's lang team meeting (notes).

Should format_args! invocations that were generated by a proc-macro be able to use implicit argument captures?

Several of us had the instinct that the answer should be "yes, if the span attached to the {foo} has hygiene that can name foo". It's possible that this answer misses some nuance, but if it does we're missing that context.

It's in my opinion unreasonable to support the indoc usage but not my example above (since only the spans of the string literal are different).

I think I agree, but I don't have context on what the spans of the string literal produced by indoc are. I would guess that it uses Span::call_site()?

indoc (or other crates like it) could do their own parsing of the implicit captures and expand them out into separate bindings before forwarding them to format_args, but this can of course be said for the whole format_args machinery.

I don't follow. Can you please give an example?

Generally, we'd like to see an explainer that compares the behavior in the RFC (to the extent that it's specified) to the old behavior, current behavior, and behavior after landing this PR.

In the short term, we think it makes sense to revert #106195 (and backport that revert if needed) so that we don't break existing code in the next release. Fixing an ICE is lower priority than keeping existing code working. Then we can discuss what the right behavior is and implement that without the time pressure.

Please re-nominate for discussion when our input is needed again. Thanks!

…ther-it-really-is-a-literal, r=compiler-errors Revert "Improve heuristics whether `format_args` string is a source literal" This reverts commit e6c02aa (from rust-lang#106195). Keeps the code improvements from the PR and the test (as a known-bug). Works around rust-lang#106408 while a proper fix is discussed more thoroughly in rust-lang#106505, as proposed by `@tmandry.` Reopens rust-lang#106191 r? compiler-errors

Nilstrieb · 2023-01-24T18:50:23Z

I reverted it in the meantime which makes the issue about indoc work again.
My last comment was a little confusing, I'll try to make it clearer now^^
FWIW I think that this is a clear bug that should be fixed like in this PR, but I want to make sure that you agree with this new behavior since it makes more code compile which we won't be able to undo in the future.

I made a small table of things macros were able to expand to before, after my PR that fixed an ICE and caused a regression and was then reverted, and this PR which fixes it all.

Macro expanding to	Originally	After my first PR	With this PR
`format!("{x}", x = 0)`	✅	✅	✅
`format!("{x}")`, where `"{x}"` is the macro input and there is an `x` ident with the same hygiene as the literal and the literal span is the same span as in the source	✅	✅	✅
`{ let x = 5; format!("{x}") }` (no macro input, any hygiene for `x`)	❌	❌	✅
`format!("A{x}")`, where `"{x}"` is the macro input and there is an `x` ident with the same hygiene as the literal. The literal has been slightly modified by the macro so its span does not point at the exact same characters in the source.	✅	❌#106408	✅
A macro that expands to a `format` invocation and modifies a non-ascii string literal in just the right way	🧊 (ICE)	✅	✅

scottmcm · 2023-01-24T19:52:17Z

We discussed this in the lang team meeting today. Broadly we thought that proc macros being able to emit tokens that use captures makes logical sense. Basically, a proc macro emitting something as a whole that one could write directly, as in https://github.com/rust-lang/rust/pull/106505/files#diff-868206cf54a88d81468a76d45e639cea40c537f2031d19f06c114e3cb329a426R44, seems entirely fine, though mixing things like format!(concat!(…)) might still be best to reject.

So I'll start a

@rfcbot fcp merge

but we had some requests to confirm, so...

rfcbot · 2023-01-24T19:52:18Z

Team member @scottmcm has proposed to merge this. The next step is review by the rest of the tagged team members:

Concerns:

~~please-link-tests~~ resolved by Properly allow macro expanded format_args invocations to uses captures #106505 (comment)

Once a majority of reviewers approve (and at most 2 approvals are outstanding), this will enter its final comment period. If you spot a major issue that hasn't been raised at any point in this process, please speak up!

cc @rust-lang/lang-advisors: FCP proposed for lang, please feel free to register concerns.
See this document for info about what commands tagged team members can give me.

scottmcm · 2023-01-24T20:02:19Z

@rfcbot concern please-link-tests

We did discuss a bunch of situations in the meeting where we'd like to get them listed out to affirmatively confirm those cases. It sounds like a bunch of them already work as expected, but we'd like to have tests (or pointers to existing tests) for things like how hygiene works in macro-generated format strings with captures, which things don't work, how it relates to when things are expanded, etc. (Like is format!(macro_be!()) different from format!(p_macro!())?)

petrochenkov · 2023-01-25T08:42:39Z

a proc macro emitting something as a whole that one could write directly ... seems entirely fine, though mixing things like format!(concat!(…)) might still be best to reject

This was always my understanding as well.
We parse the expression at the start of TOKENS in format_args!(TOKENS), if it's a literal - good, if it's not a literal - bad.
It's not clear why the implementation considered spans at all.

petrochenkov · 2023-01-25T08:52:15Z

The current implementation seems overcomplicated too - we don't need is_string_literal_including_nt, we only need to parse an expression and then do a post factum check on it.

tmandry · 2023-01-26T01:24:19Z

{ let x = 5; format!("{x}") } (no macro input, any hygiene for x)

This PR still requires the hygiene for x in the binding and format string to match, right?

Nilstrieb · 2023-01-26T06:31:33Z

Yes. Well, it doesn't require them to match for macro expansion to succeed, but x will not be found later if it doesn't match.

Nilstrieb · 2023-01-29T19:41:25Z

I have collected and added a few tests:

tests/ui/fmt/indoc-issue-106408.rs tests that hygiene info is preserved correctly through a proc macro invocation
tests/ui/fmt/format-args-capture-first-literal-is-macro.rs tests that MBEs, proc macros and builtin macros in argument position block captures
tests/ui/fmt/format-args-capture-from-pm-first-arg-macro.rs tests that we still forbid format!(concat!("{a}")) even if it comes from a proc macro

joshtriplett · 2023-01-31T19:18:28Z

@scottmcm Do the latest changes by @Nilstrieb address your concern?

scottmcm · 2023-02-03T21:23:36Z

Thanks, @Nilstrieb ! That's very helpful
@rfcbot resolve please-link-tests

The only other thing that comes to mind is something like this

macro_rules! format_mbe {
    ($tt:tt) => {
        {
            let a = 123;
            format!($tt)
        }
    };
}

fn main() {
    let a = 0;
    assert_eq!(format_mbe!("{a}"), "0");
}

but if that's not actually different because of how resolution works, then feel free to omit it.

compiler/rustc_parse_format/src/lib.rs

petrochenkov · 2023-03-14T13:12:33Z

r=me after addressing the remaining comments.
@rustbot author

Originally, this was kinda half-allowed. There were some primitive checks in place that looked at the span to see whether the input was likely a literal. These "source literal" checks are needed because the spans created during `format_args` parsing only make sense when it is indeed a literal that was written in the source code directly. This is orthogonal to the restriction that the first argument must be a "direct literal", not being exanpanded from macros. This restriction was imposed by [RFC 2795] on the basis of being too confusing. But this was only concerned with the argument of the invocation being a literal, not whether it was a source literal (maybe in spirit it meant it being a source literal, this is not clear to me). Since the original check only really cared about source literals (which is good enough to deny the `format_args!(concat!())` example), macros expanding to `format_args` invocations were able to use implicit captures if they spanned the string in a way that lead back to a source string. The "source literal" checks were not strict enough and caused ICEs in certain cases (see # 106191 (the space is intended to avoid spammy backreferences)). So I tightened it up in # 106195 to really only work if it's a direct source literal. This caused the `indoc` crate to break. `indoc` transformed the source literal by removing whitespace, which made it not a "source literal" anymore (which is required to fix the ICE). But since `indoc` spanned the literal in ways that made the old check think that it's a literal, it was able to use implicit captures (which is useful and nice for the users of `indoc`). This commit properly seperates the previously introduced concepts of "source literal" and "direct literal" and therefore allows `indoc` invocations, which don't create "source literals" to use implicit captures again. [RFC 2795]: https://rust-lang.github.io/rfcs/2795-format-args-implicit-identifiers.html#macro-hygiene

Sometimes, we want to create subspans and point at code in the literal if possible. But this doesn't always make sense, sometimes the literal may come from macro expanded code and isn't actually there in the source. Then, we can't really make these suggestions. This now makes sure that the literal is actually there as we see it so that we will not run into ICEs on weird literal transformations.

Nilstrieb · 2023-03-14T14:16:28Z

@bors r=petrochenkov

bors · 2023-03-14T14:16:30Z

📌 Commit 427aceb has been approved by petrochenkov

It is now in the queue for this repository.

bors · 2023-03-14T14:25:05Z

⌛ Testing commit 427aceb with merge 2e7034e...

bors · 2023-03-14T17:40:44Z

☀️ Test successful - checks-actions
Approved by: petrochenkov
Pushing 2e7034e to master...

rust-timer · 2023-03-14T19:29:59Z

Finished benchmarking commit (2e7034e): comparison URL.

Overall result: ❌ regressions - no action needed

@rustbot label: -perf-regression

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	0.4%	[0.4%, 0.5%]	5
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-	-	0

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)	3.5%	[3.5%, 3.5%]	1
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-3.5%	[-3.5%, -3.5%]	1
All ❌✅ (primary)	3.5%	[3.5%, 3.5%]	1

Cycles

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
Regressions ❌ (secondary)	4.9%	[4.1%, 5.5%]	4
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.2%	[-2.3%, -2.1%]	2
All ❌✅ (primary)	-	-	0

Pkgsrc changes: * Adjust patches and cargo checksums to new versions. * Adjust to not cross-build to 8.0, due to LLVM using c++17, so adjust USE_LANGUAGES. Upstream changes: Version 1.70.0 (2023-06-01) ========================== Language -------- - [Relax ordering rules for `asm!` operands] (rust-lang/rust#105798) - [Properly allow macro expanded `format_args` invocations to uses captures] (rust-lang/rust#106505) - [Lint ambiguous glob re-exports] (rust-lang/rust#107880) - [Perform const and unsafe checking for expressions in `let _ = expr` position.] (rust-lang/rust#102256) Compiler -------- - [Extend -Cdebuginfo with new options and named aliases] (rust-lang/rust#109808) This provides a smaller version of debuginfo for cases that only need line number information (`-Cdebuginfo=line-tables-only`), which may eventually become the default for `-Cdebuginfo=1`. - [Make `unused_allocation` lint against `Box::new` too] (rust-lang/rust#104363) - [Detect uninhabited types early in const eval] (rust-lang/rust#109435) - [Switch to LLD as default linker for {arm,thumb}v4t-none-eabi] (rust-lang/rust#109721) - [Add tier 3 target `loongarch64-unknown-linux-gnu`] (rust-lang/rust#96971) - [Add tier 3 target for `i586-pc-nto-qnx700` (QNX Neutrino RTOS, version 7.0)] (rust-lang/rust#109173), - [Insert alignment checks for pointer dereferences as debug assertions] (rust-lang/rust#98112) This catches undefined behavior at runtime, and may cause existing code to fail. Refer to Rust's [platform support page][platform-support-doc] for more information on Rust's tiered platform support. Libraries --------- - [Document NonZeroXxx layout guarantees] (rust-lang/rust#94786) - [Windows: make `Command` prefer non-verbatim paths] (rust-lang/rust#96391) - [Implement Default for some alloc/core iterators] (rust-lang/rust#99929) - [Fix handling of trailing bare CR in str::lines] (rust-lang/rust#100311) - [allow negative numeric literals in `concat!`] (rust-lang/rust#106844) - [Add documentation about the memory layout of `Cell`] (rust-lang/rust#106921) - [Use `partial_cmp` to implement tuple `lt`/`le`/`ge`/`gt`] (rust-lang/rust#108157) - [Stabilize `atomic_as_ptr`] (rust-lang/rust#108419) - [Stabilize `nonnull_slice_from_raw_parts`] (rust-lang/rust#97506) - [Partial stabilization of `once_cell`] (rust-lang/rust#105587) - [Stabilize `nonzero_min_max`] (rust-lang/rust#106633) - [Flatten/inline format_args!() and (string and int) literal arguments into format_args!()] (rust-lang/rust#106824) - [Stabilize movbe target feature] (rust-lang/rust#107711) - [don't splice from files into pipes in io::copy] (rust-lang/rust#108283) - [Add a builtin unstable `FnPtr` trait that is implemented for all function pointers] (rust-lang/rust#108080) This extends `Debug`, `Pointer`, `Hash`, `PartialEq`, `Eq`, `PartialOrd`, and `Ord` implementations for function pointers with all ABIs. Stabilized APIs --------------- - [`NonZero*::MIN/MAX`] (https://doc.rust-lang.org/stable/std/num/struct.NonZeroI8.html#associatedconstant.MIN) - [`BinaryHeap::retain`] (https://doc.rust-lang.org/stable/std/collections/struct.BinaryHeap.html#method.retain) - [`Default for std::collections::binary_heap::IntoIter`] (https://doc.rust-lang.org/stable/std/collections/binary_heap/struct.IntoIter.html) - [`Default for std::collections::btree_map::{IntoIter, Iter, IterMut}`] (https://doc.rust-lang.org/stable/std/collections/btree_map/struct.IntoIter.html) - [`Default for std::collections::btree_map::{IntoKeys, Keys}`] (https://doc.rust-lang.org/stable/std/collections/btree_map/struct.IntoKeys.html) - [`Default for std::collections::btree_map::{IntoValues, Values}`] (https://doc.rust-lang.org/stable/std/collections/btree_map/struct.IntoKeys.html) - [`Default for std::collections::btree_map::Range`] (https://doc.rust-lang.org/stable/std/collections/btree_map/struct.Range.html) - [`Default for std::collections::btree_set::{IntoIter, Iter}`] (https://doc.rust-lang.org/stable/std/collections/btree_set/struct.IntoIter.html) - [`Default for std::collections::btree_set::Range`] (https://doc.rust-lang.org/stable/std/collections/btree_set/struct.Range.html) - [`Default for std::collections::linked_list::{IntoIter, Iter, IterMut}`] (https://doc.rust-lang.org/stable/alloc/collections/linked_list/struct.IntoIter.html) - [`Default for std::vec::IntoIter`] (https://doc.rust-lang.org/stable/alloc/vec/struct.IntoIter.html#impl-Default-for-IntoIter%3CT,+A%3E) - [`Default for std::iter::Chain`] (https://doc.rust-lang.org/stable/std/iter/struct.Chain.html) - [`Default for std::iter::Cloned`] (https://doc.rust-lang.org/stable/std/iter/struct.Cloned.html) - [`Default for std::iter::Copied`] (https://doc.rust-lang.org/stable/std/iter/struct.Copied.html) - [`Default for std::iter::Enumerate`] (https://doc.rust-lang.org/stable/std/iter/struct.Enumerate.html) - [`Default for std::iter::Flatten`] (https://doc.rust-lang.org/stable/std/iter/struct.Flatten.html) - [`Default for std::iter::Fuse`] (https://doc.rust-lang.org/stable/std/iter/struct.Fuse.html) - [`Default for std::iter::Rev`] (https://doc.rust-lang.org/stable/std/iter/struct.Rev.html) - [`Default for std::slice::Iter`] (https://doc.rust-lang.org/stable/std/slice/struct.Iter.html) - [`Default for std::slice::IterMut`] (https://doc.rust-lang.org/stable/std/slice/struct.IterMut.html) - [`Rc::into_inner`] (https://doc.rust-lang.org/stable/alloc/rc/struct.Rc.html#method.into_inner) - [`Arc::into_inner`] (https://doc.rust-lang.org/stable/alloc/sync/struct.Arc.html#method.into_inner) - [`std::cell::OnceCell`] (https://doc.rust-lang.org/stable/std/cell/struct.OnceCell.html) - [`Option::is_some_and`] (https://doc.rust-lang.org/stable/std/option/enum.Option.html#method.is_some_and) - [`NonNull::slice_from_raw_parts`] (https://doc.rust-lang.org/stable/std/ptr/struct.NonNull.html#method.slice_from_raw_parts) - [`Result::is_ok_and`] (https://doc.rust-lang.org/stable/std/result/enum.Result.html#method.is_ok_and) - [`Result::is_err_and`] (https://doc.rust-lang.org/stable/std/result/enum.Result.html#method.is_err_and) - [`std::sync::atomic::Atomic*::as_ptr`] (https://doc.rust-lang.org/stable/std/sync/atomic/struct.AtomicU8.html#method.as_ptr) - [`std::io::IsTerminal`] (https://doc.rust-lang.org/stable/std/io/trait.IsTerminal.html) - [`std::os::linux::net::SocketAddrExt`] (https://doc.rust-lang.org/stable/std/os/linux/net/trait.SocketAddrExt.html) - [`std::os::unix::net::UnixDatagram::bind_addr`] (https://doc.rust-lang.org/stable/std/os/unix/net/struct.UnixDatagram.html#method.bind_addr) - [`std::os::unix::net::UnixDatagram::connect_addr`] (https://doc.rust-lang.org/stable/std/os/unix/net/struct.UnixDatagram.html#method.connect_addr) - [`std::os::unix::net::UnixDatagram::send_to_addr`] (https://doc.rust-lang.org/stable/std/os/unix/net/struct.UnixDatagram.html#method.send_to_addr) - [`std::os::unix::net::UnixListener::bind_addr`] (https://doc.rust-lang.org/stable/std/os/unix/net/struct.UnixListener.html#method.bind_addr) - [`std::path::Path::as_mut_os_str`] (https://doc.rust-lang.org/stable/std/path/struct.Path.html#method.as_mut_os_str) - [`std::sync::OnceLock`] (https://doc.rust-lang.org/stable/std/sync/struct.OnceLock.html) Cargo ----- - [Add `CARGO_PKG_README`] (rust-lang/cargo#11645) - [Make `sparse` the default protocol for crates.io] (rust-lang/cargo#11791) - [Accurately show status when downgrading dependencies] (rust-lang/cargo#11839) - [Use registry.default for login/logout] (rust-lang/cargo#11949) - [Stabilize `cargo logout`] (rust-lang/cargo#11950) Misc ---- - [Stabilize rustdoc `--test-run-directory`] (rust-lang/rust#103682) Compatibility Notes ------------------- - [Prevent stable `libtest` from supporting `-Zunstable-options`] (rust-lang/rust#109044) - [Perform const and unsafe checking for expressions in `let _ = expr` position.] (rust-lang/rust#102256) - [WebAssembly targets enable `sign-ext` and `mutable-globals` features in codegen] (rust-lang/rust#109807) This may cause incompatibility with older execution environments. - [Insert alignment checks for pointer dereferences as debug assertions] (rust-lang/rust#98112) This catches undefined behavior at runtime, and may cause existing code to fail. 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. - [Upgrade to LLVM 16] (rust-lang/rust#109474) - [Use SipHash-1-3 instead of SipHash-2-4 for StableHasher] (rust-lang/rust#107925)

Pkgsrc changes: * Adjust patches and cargo checksums to new versions. * Add support for NetBSD/riscv64. Upstream changes: Version 1.70.0 (2023-06-01) ========================== Language -------- - [Relax ordering rules for `asm!` operands] (rust-lang/rust#105798) - [Properly allow macro expanded `format_args` invocations to uses captures] (rust-lang/rust#106505) - [Lint ambiguous glob re-exports] (rust-lang/rust#107880) - [Perform const and unsafe checking for expressions in `let _ = expr` position.] (rust-lang/rust#102256) Compiler -------- - [Extend -Cdebuginfo with new options and named aliases] (rust-lang/rust#109808) This provides a smaller version of debuginfo for cases that only need line number information (`-Cdebuginfo=line-tables-only`), which may eventually become the default for `-Cdebuginfo=1`. - [Make `unused_allocation` lint against `Box::new` too] (rust-lang/rust#104363) - [Detect uninhabited types early in const eval] (rust-lang/rust#109435) - [Switch to LLD as default linker for {arm,thumb}v4t-none-eabi] (rust-lang/rust#109721) - [Add tier 3 target `loongarch64-unknown-linux-gnu`] (rust-lang/rust#96971) - [Add tier 3 target for `i586-pc-nto-qnx700`(QNX Neutrino RTOS, version 7.0)] (rust-lang/rust#109173), - [Insert alignment checks for pointer dereferences as debug assertions] (rust-lang/rust#98112) This catches undefined behavior at runtime, and may cause existing code to fail. Refer to Rust's [platform support page][platform-support-doc] for more information on Rust's tiered platform support. Libraries --------- - [Document NonZeroXxx layout guarantees] (rust-lang/rust#94786) - [Windows: make `Command` prefer non-verbatim paths] (rust-lang/rust#96391) - [Implement Default for some alloc/core iterators] (rust-lang/rust#99929) - [Fix handling of trailing bare CR in str::lines] (rust-lang/rust#100311) - [allow negative numeric literals in `concat!`] (rust-lang/rust#106844) - [Add documentation about the memory layout of `Cell`] (rust-lang/rust#106921) - [Use `partial_cmp` to implement tuple `lt`/`le`/`ge`/`gt`] (rust-lang/rust#108157) - [Stabilize `atomic_as_ptr`] (rust-lang/rust#108419) - [Stabilize `nonnull_slice_from_raw_parts`] (rust-lang/rust#97506) - [Partial stabilization of `once_cell`] (rust-lang/rust#105587) - [Stabilize `nonzero_min_max`] (rust-lang/rust#106633) - [Flatten/inline format_args!() and (string and int) literal arguments into format_args!()] (rust-lang/rust#106824) - [Stabilize movbe target feature] (rust-lang/rust#107711) - [don't splice from files into pipes in io::copy] (rust-lang/rust#108283) - [Add a builtin unstable `FnPtr` trait that is implemented for all function pointers] (rust-lang/rust#108080) This extends `Debug`, `Pointer`, `Hash`, `PartialEq`, `Eq`, `PartialOrd`, and `Ord` implementations for function pointers with all ABIs. Stabilized APIs --------------- - [`NonZero*::MIN/MAX`] (https://doc.rust-lang.org/stable/std/num/struct.NonZeroI8.html#associatedconstant.MIN) - [`BinaryHeap::retain`] (https://doc.rust-lang.org/stable/std/collections/struct.BinaryHeap.html#method.retain) - [`Default for std::collections::binary_heap::IntoIter`] (https://doc.rust-lang.org/stable/std/collections/binary_heap/struct.IntoIter.html) - [`Default for std::collections::btree_map::{IntoIter, Iter, IterMut}`] (https://doc.rust-lang.org/stable/std/collections/btree_map/struct.IntoIter.html) - [`Default for std::collections::btree_map::{IntoKeys, Keys}`] (https://doc.rust-lang.org/stable/std/collections/btree_map/struct.IntoKeys.html) - [`Default for std::collections::btree_map::{IntoValues, Values}`] (https://doc.rust-lang.org/stable/std/collections/btree_map/struct.IntoKeys.html) - [`Default for std::collections::btree_map::Range`] (https://doc.rust-lang.org/stable/std/collections/btree_map/struct.Range.html) - [`Default for std::collections::btree_set::{IntoIter, Iter}`] (https://doc.rust-lang.org/stable/std/collections/btree_set/struct.IntoIter.html) - [`Default for std::collections::btree_set::Range`] (https://doc.rust-lang.org/stable/std/collections/btree_set/struct.Range.html) - [`Default for std::collections::linked_list::{IntoIter, Iter, IterMut}`] (https://doc.rust-lang.org/stable/alloc/collections/linked_list/struct.IntoIter.html) - [`Default for std::vec::IntoIter`] (https://doc.rust-lang.org/stable/alloc/vec/struct.IntoIter.html#impl-Default-for-IntoIter%3CT,+A%3E) - [`Default for std::iter::Chain`] (https://doc.rust-lang.org/stable/std/iter/struct.Chain.html) - [`Default for std::iter::Cloned`] (https://doc.rust-lang.org/stable/std/iter/struct.Cloned.html) - [`Default for std::iter::Copied`] (https://doc.rust-lang.org/stable/std/iter/struct.Copied.html) - [`Default for std::iter::Enumerate`] (https://doc.rust-lang.org/stable/std/iter/struct.Enumerate.html) - [`Default for std::iter::Flatten`] (https://doc.rust-lang.org/stable/std/iter/struct.Flatten.html) - [`Default for std::iter::Fuse`] (https://doc.rust-lang.org/stable/std/iter/struct.Fuse.html) - [`Default for std::iter::Rev`] (https://doc.rust-lang.org/stable/std/iter/struct.Rev.html) - [`Default for std::slice::Iter`] (https://doc.rust-lang.org/stable/std/slice/struct.Iter.html) - [`Default for std::slice::IterMut`] (https://doc.rust-lang.org/stable/std/slice/struct.IterMut.html) - [`Rc::into_inner`] (https://doc.rust-lang.org/stable/alloc/rc/struct.Rc.html#method.into_inner) - [`Arc::into_inner`] (https://doc.rust-lang.org/stable/alloc/sync/struct.Arc.html#method.into_inner) - [`std::cell::OnceCell`] (https://doc.rust-lang.org/stable/std/cell/struct.OnceCell.html) - [`Option::is_some_and`] (https://doc.rust-lang.org/stable/std/option/enum.Option.html#method.is_some_and) - [`NonNull::slice_from_raw_parts`] (https://doc.rust-lang.org/stable/std/ptr/struct.NonNull.html#method.slice_from_raw_parts) - [`Result::is_ok_and`] (https://doc.rust-lang.org/stable/std/result/enum.Result.html#method.is_ok_and) - [`Result::is_err_and`] (https://doc.rust-lang.org/stable/std/result/enum.Result.html#method.is_err_and) - [`std::sync::atomic::Atomic*::as_ptr`] (https://doc.rust-lang.org/stable/std/sync/atomic/struct.AtomicU8.html#method.as_ptr) - [`std::io::IsTerminal`] (https://doc.rust-lang.org/stable/std/io/trait.IsTerminal.html) - [`std::os::linux::net::SocketAddrExt`] (https://doc.rust-lang.org/stable/std/os/linux/net/trait.SocketAddrExt.html) - [`std::os::unix::net::UnixDatagram::bind_addr`] (https://doc.rust-lang.org/stable/std/os/unix/net/struct.UnixDatagram.html#method.bind_addr) - [`std::os::unix::net::UnixDatagram::connect_addr`] (https://doc.rust-lang.org/stable/std/os/unix/net/struct.UnixDatagram.html#method.connect_addr) - [`std::os::unix::net::UnixDatagram::send_to_addr`] (https://doc.rust-lang.org/stable/std/os/unix/net/struct.UnixDatagram.html#method.send_to_addr) - [`std::os::unix::net::UnixListener::bind_addr`] (https://doc.rust-lang.org/stable/std/os/unix/net/struct.UnixListener.html#method.bind_addr) - [`std::path::Path::as_mut_os_str`] (https://doc.rust-lang.org/stable/std/path/struct.Path.html#method.as_mut_os_str) - [`std::sync::OnceLock`] (https://doc.rust-lang.org/stable/std/sync/struct.OnceLock.html) Cargo ----- - [Add `CARGO_PKG_README`] (rust-lang/cargo#11645) - [Make `sparse` the default protocol for crates.io] (rust-lang/cargo#11791) - [Accurately show status when downgrading dependencies] (rust-lang/cargo#11839) - [Use registry.default for login/logout] (rust-lang/cargo#11949) - [Stabilize `cargo logout`] (rust-lang/cargo#11950) Misc ---- - [Stabilize rustdoc `--test-run-directory`] (rust-lang/rust#103682) Compatibility Notes ------------------- - [Prevent stable `libtest` from supporting `-Zunstable-options`] (rust-lang/rust#109044) - [Perform const and unsafe checking for expressions in `let _ = expr` position.] (rust-lang/rust#102256) - [WebAssembly targets enable `sign-ext` and `mutable-globals` features in codegen] (rust-lang/rust#109807) This may cause incompatibility with older execution environments. - [Insert alignment checks for pointer dereferences as debug assertions] (rust-lang/rust#98112) This catches undefined behavior at runtime, and may cause existing code to fail. 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. - [Upgrade to LLVM 16] (rust-lang/rust#109474) - [Use SipHash-1-3 instead of SipHash-2-4 for StableHasher] (rust-lang/rust#107925)

rustbot added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. T-compiler Relevant to the compiler team, which will review and decide on the PR/issue. labels Jan 5, 2023

Nilstrieb added the I-lang-nominated The issue / PR has been nominated for discussion during a lang team meeting. label Jan 5, 2023

tmandry removed the I-lang-nominated The issue / PR has been nominated for discussion during a lang team meeting. label Jan 11, 2023

Nilstrieb mentioned this pull request Jan 18, 2023

Revert "Improve heuristics whether format_args string is a source literal" #107041

Merged

Nilstrieb added the I-lang-nominated The issue / PR has been nominated for discussion during a lang team meeting. label Jan 24, 2023

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 Jan 24, 2023

petrochenkov self-assigned this Jan 25, 2023

petrochenkov added S-waiting-on-team Status: Awaiting decision from the relevant subteam (see the T-<team> label). and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Jan 25, 2023

Nilstrieb force-pushed the format-args-string-literal-episode-2 branch from c3e168b to 5f2fb28 Compare January 29, 2023 19:00

Nilstrieb force-pushed the format-args-string-literal-episode-2 branch from 5f2fb28 to ed966c4 Compare January 29, 2023 19:44

petrochenkov added the S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. label Mar 13, 2023

rust-cloud-vms bot force-pushed the format-args-string-literal-episode-2 branch 2 times, most recently from 9c2894d to eae116f Compare March 13, 2023 18:55

Nilstrieb added S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. and removed S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Mar 13, 2023

petrochenkov reviewed Mar 14, 2023

View reviewed changes

compiler/rustc_parse_format/src/lib.rs Outdated Show resolved Hide resolved

compiler/rustc_parse_format/src/lib.rs Outdated Show resolved Hide resolved

rustbot added S-waiting-on-author Status: This is awaiting some action (such as code changes or more information) from the author. and removed S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. labels Mar 14, 2023

Nilstrieb added 2 commits March 14, 2023 13:16

rust-cloud-vms bot force-pushed the format-args-string-literal-episode-2 branch from eae116f to 427aceb Compare March 14, 2023 13:20

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-author Status: This is awaiting some action (such as code changes or more information) from the author. labels Mar 14, 2023

bors mentioned this pull request Mar 14, 2023

remove obsolete givens from regionck #107376

Merged

bors added the merged-by-bors This PR was explicitly merged by bors. label Mar 14, 2023

bors merged commit 2e7034e into rust-lang:master Mar 14, 2023
11 checks passed

rustbot added this to the 1.70.0 milestone Mar 14, 2023

apiraino removed the to-announce Announce this issue on triage meeting label Mar 23, 2023

Nilstrieb mentioned this pull request Apr 27, 2023

[DO NOT MERGE] have tag be explicit debug_assert! argument #110888

Closed

pravic mentioned this pull request May 29, 2023

Github rejects rendering of RELEASES.md #112058

Closed

Nilstrieb deleted the format-args-string-literal-episode-2 branch November 5, 2023 07:56

matthiaskrgr mentioned this pull request Nov 7, 2023

ice: const_trait_impl: async functions cannot be const fn #117629

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Properly allow macro expanded `format_args` invocations to uses captures #106505

Properly allow macro expanded `format_args` invocations to uses captures #106505

Nilstrieb commented Jan 5, 2023 •

edited

Nilstrieb commented Jan 5, 2023 •

edited

tmandry commented Jan 11, 2023 •

edited

Nilstrieb commented Jan 24, 2023

scottmcm commented Jan 24, 2023 •

edited

rfcbot commented Jan 24, 2023 •

edited by tmandry

scottmcm commented Jan 24, 2023

petrochenkov commented Jan 25, 2023

petrochenkov commented Jan 25, 2023

tmandry commented Jan 26, 2023

Nilstrieb commented Jan 26, 2023

Nilstrieb commented Jan 29, 2023

joshtriplett commented Jan 31, 2023

scottmcm commented Feb 3, 2023

petrochenkov commented Mar 14, 2023

Nilstrieb commented Mar 14, 2023

bors commented Mar 14, 2023

bors commented Mar 14, 2023

bors commented Mar 14, 2023

rust-timer commented Mar 14, 2023

Properly allow macro expanded format_args invocations to uses captures #106505

Properly allow macro expanded format_args invocations to uses captures #106505

Conversation

Nilstrieb commented Jan 5, 2023 • edited

Nilstrieb commented Jan 5, 2023 • edited

Question for the lang team

Arguments for it

Arguments against it

tmandry commented Jan 11, 2023 • edited

Nilstrieb commented Jan 24, 2023

scottmcm commented Jan 24, 2023 • edited

rfcbot commented Jan 24, 2023 • edited by tmandry

scottmcm commented Jan 24, 2023

petrochenkov commented Jan 25, 2023

petrochenkov commented Jan 25, 2023

tmandry commented Jan 26, 2023

Nilstrieb commented Jan 26, 2023

Nilstrieb commented Jan 29, 2023

joshtriplett commented Jan 31, 2023

scottmcm commented Feb 3, 2023

petrochenkov commented Mar 14, 2023

Nilstrieb commented Mar 14, 2023

bors commented Mar 14, 2023

bors commented Mar 14, 2023

bors commented Mar 14, 2023

rust-timer commented Mar 14, 2023

Overall result: ❌ regressions - no action needed

Instruction count

Max RSS (memory usage)

Cycles

Properly allow macro expanded `format_args` invocations to uses captures #106505

Properly allow macro expanded `format_args` invocations to uses captures #106505

Nilstrieb commented Jan 5, 2023 •

edited

Nilstrieb commented Jan 5, 2023 •

edited

tmandry commented Jan 11, 2023 •

edited

scottmcm commented Jan 24, 2023 •

edited

rfcbot commented Jan 24, 2023 •

edited by tmandry