August 2023 Monthly report

Overview

Thanks again to Open Source Security, inc and Embecosm for their ongoing support for this project.

Milestone Progress

Another very important month in terms of contributions, with 86 pull requests merged. Once again, most of them come from the Google Summer of Code program, which is coming to an end. You can expect more detailed blogposts about our students’ work this summer in the coming weeks, but the short of it is: they have both been hugely successful, and have proved to be amazing contributors to the project. They have both passed the GSoC evaluations brillantly, and we are hoping they will keep working on the project! Thank you Raiki Tamura and Mahad Muhammad! Well done!

On the technical side of things, late name resolution is going swimmingly. We are now in the process of merging it in the compiler, while still keeping it hidden behind a specific command line flag (-frust-name-resolution-2.0). A lot of our existing testsuite passes with that new option enabled, and we are working on fixing the remaining failures. Pierre- Emmanuel has done some really good work on our early name resolution too, taking care of transforming it into a fixed point algorithm which runs until nothing has changed. This enables interesting behaviors, such as a macro invocation expanding to a macro definition, which a later invocation could resolve to. Early name resolution is also responsible for handling imports (use declarations), and we are on track to add the last piece of the puzzle needed for resolving all core 1.49 imports: glob imports (use foo::*). These imports work differently from regular imports, as they only concern public items, and can be shadowed by later declarations, but they are much easier to handle in our new name resolution algorithm and are almost there.

This resolution pass also needs to tie in with the rest of the codebase, namely typechecking and our compilation backend, which rely on it closely. We have spent a long amount of time recreating the missing functions from those new data structures, and making sure the later compilation passes could make use of the new resolution.

On the typechecking side, Philip is also getting infuriatingly close to getting iterators working! Thanks to advice from bjorn3 and Matthew Jasper, we have identified a few bugs in our handling of compiler intrinsics, which are now fixed. The handling of opaque types is one of the last big pieces of iterator support, which Philip is currently working on.

After that work is complete, we will start working on the format_args!() builtin macro, a central component of the widely used println!() macro as well as a lot of print-like macros available in Linux kernel.

macro_rules! print_macro (
    // The non-continuation cases (most of them, e.g. `INFO`).
    ($format_string:path, false, $($arg:tt)+) => (
        // To remain sound, `arg`s must be expanded outside the `unsafe` block.
        // Typically one would use a `let` binding for that; however, `format_args!`
        // takes borrows on the arguments, but does not extend the scope of temporaries.
        // Therefore, a `match` expression is used to keep them around, since
        // the scrutinee is kept until the end of the `match`.
        match format_args!($($arg)+) {
            // SAFETY: This hidden macro should only be called by the documented
            // printing macros which ensure the format string is one of the fixed
            // ones. All `__LOG_PREFIX`s are null-terminated as they are generated
            // by the `module!` proc macro or fixed values defined in a kernel
            // crate.
            args => unsafe {
                $crate::print::call_printk(
                    &$format_string,
                    crate::__LOG_PREFIX,
                    args,
                );
            }
        }
    );

    // The `CONT` case.
    ($format_string:path, true, $($arg:tt)+) => (
        $crate::print::call_printk_cont(
            format_args!($($arg)+),
        );
    );
);

Source: https://github.com/Rust-for-Linux/linux/blob/4af84c6a85c63bec24611e46bb3de2c0a6602a51/rust/kernel/print.rs#L146-L176.

Finally, we are looking forward to meeting with the GCC community once again during GNU Cauldron at the end of September in Cambridge. Philip and Pierre-Emmanuel will be giving a talk about some of the recent changes in the compiler, as well as the inner workings of procedural macros. We will also be talking about the compilation of the core library at EuroRust in Brussels. See you then!

Community call

We will have our next monthly community call on the 11th of September

1. You can subscribe to our calendar to see when the next one will be held. The call is open to everyone, even if you would just like to sit-in and listen. You can also subscribe to our mailing-list or join our Zulip chat to be notified of upcoming events.

• jitsi link
• Calendar ID: 7060a0923ffebd3cb52b1afef35a28ff7b64f05962c9af84c23b1847f1f5f894@group.calendar.google.com
• Google calendar link
• iCal link

Completed Activities

• Move statement-related methods into base class ```Backend``` PR2602
• Move function-related methods into base class ```Backend``` PR2601
• Move some simple expression-related methods into base class ```Backend``` PR2600
• Move variable-related methods into base class ```Backend``` PR2599
• [E0603] Use of private item outside scope PR2597
• Improve type checking for if expressions PR2596
• Fix ODR violations PR2595
• Bump our docker build’s debian base image version to match ```rust:latest```’s PR2594
• Move label-related functions into base class ```Backend``` PR2593
• Move debug-related functions into base class ```Backend``` PR2592
• Bump ```docker/build-push-action``` to v4 PR2590
• Move type-related functions into base class ```Backend``` PR2589
• Bump workfow versions PR2588
• Handle remaining unhandled gengtype annotations PR2586
• Fix iterator intrinsics PR2584
• Add missing ```gengtype``` related structs PR2580
• Rename some structs to avoid ODR issues PR2579
• Fix invalid call to ```vector::front``` in ```TypeBoundPredicate``` constructors PR2577
• gccrs: Fix compilation of types which hold onto dangling infer vars PR2575
• diagnostics: Added non-const `rich<sub>location</sub> *` function PR2574
• Remove ```stmt<sub>tree</sub>``` and dependency on ```cp<sub>tokencache</sub>``` PR2573
• Namespace related tweaks PR2572
• Fork ```c++``` resorting methods PR2570
• Top level or-patterns are not allowed let binding PR2569
• [E0573] Something other than type was used PR2568
• [E0164] Neither tuple struct nor tuple variant used as a pattern PR2565
• [E0769] Use of struct or tuple variant in struct or struct variant PR2564
• [E0532] Pattern arm did not match expected kind. PR2563
• [E0271] Type mismatch between associated type trait. PR2562
• [E0391] Detected type dependency cycle PR2561
• [E0034] found more than one items for method PR2560
• [E0535] Unknown argument given to inline attribute PR2559
• [E0534] inline attribute was malformed PR2558
• Invalid order of generic parameters PR2557
• Non-allowed default type paramters PR2556
• [E0592] method or associated functions already defined with same names PR2555
• Handle gengtype annotations in ```rust-constexpr.cc``` PR2554
• `#[no<sub>mangle</sub>]` should not be applied to non-ASCII items PR2552
• gccrs: improve name mangling hash PR2551
• gccrs: simplify matching possible candidates PR2550
• gccrs: remove horrible hack in solving complex generics on impl blocks PR2549
• gccrs: Fix legacy mangling to use Unicode escape PR2547
• Add missing ```CHECKING<sub>P</sub>``` guard to ```rust-punycode.cc``` PR2546
• gccrs: Cleanup HIR dump PR2544
• gccrs: Adjust HIR dump for IfExpr and IfExprConseqElse PR2543
• Support for `rich-loc` & `errorcode` in parser & expansion errors PR2542
• Handle nested macro definition (mappings) PR2541
• Basic formatting/comment fixes PR2540
• [E0107] Wrong number of generic argument PR2539
• [E0015] Use of non-const inside const PR2538
• [E0658] Use of mutable reference in constant functions. PR2536
• Add punycode encoding to v0 mangling PR2535
• [E0412] used type name not in scope PR2534
• Implement punycode encoding PR2533
• Fix a warning about extra semicolon PR2532
• [E0753] Use of inner doc comment in invalid context PR2531
• Move Codepoint to gcc/rust/util/ PR2530
• Add type annotation to make-rust-unicode-data.py PR2529
• `[E0054-E0604-E0620-E0606]` TypeCasting ErrorCodes PR2528
• [E0380] Use of auto trait with method or associated item PR2527
• [E0541] Use of unknown meta item PR2526
• [E0308] array misamatch types PR2525
• [E0282] type annotations needed PR2524
• macro-expand: Keep optional references to last_{def, invoc} PR2520
• [E0703] Use of Invalid ABI PR2519
• diagnostics: Fix mismatch between new[] and free PR2518
• Fix lexer exponential on tuple PR2515
• [E0635] Use of Unknown feature. PR2513
• Improve byte vector to string conversion PR2507
• Remove ```Backend::bool<sub>type</sub>``` and ```Backend::char<sub>type</sub>``` PR2506
• Move ```Backend::error<sub>variable</sub>``` to ```Bvariable::error<sub>variable</sub>``` PR2505
• mklog: handle Signed-Off-By, minor cleanup PR2504
• Remove unneeded references to ```this``` in backend methods PR2503
• [E0426] Use of undeclared label PR2502
• Use ```MacroInvocLexerBase``` in ```ProcMacroInvocLexer``` PR2498
• Use ```MacroInvocLexerBase``` in ```MacroInvocLexer``` PR2497
• [E0308] mismatch types on both sides of assignment Operator PR2494
• [E0572] return is outside of function context PR2493
• fix bootstrap build PR2492
• Normalize all identifiers to NFC PR2489
• nr2.0: Add base for Early name resolution PR2470
• nr2.0: Add DefaultVisitor class PR2469
• Normalize Hangul PR2467
• gccrs: use tl::optional for macro transcribers PR2464
• name resolution 2.0: Add base Resolver and TopLevel pass PR2456
• gccrs: better handling of incorrect file loading in include_* builtin macros PR2422

Contributors this month

• Raiki Tamura
• Mahad Muhammad
• Marc Poulhiès
• Owen Avery

Overall Task Status

Test Cases

Bugs

Milestones Progress

Testing project

Planned Activities

• Opaque types
• Fix remaining iterator bugs for core 1.49
• Merge late name resolution 2.0
• Fix remaining failing test cases under name resolution 2.0
• Drop marker trait

Detailed changelog