Document repositories maintained by the T-compiler #918

apiraino · 2025-07-17T11:27:54Z

There was a discussion on Zulip a while ago about "what happens when one becomes a team member". One of the confusing things was (at least in my experience) that one gains write permissions on a number of git repositories under the organization umbrella without being able to understand what they are.

So, following a suggestion of @jieyouxu (hey 👋 ), I thought about a page trying to add a bit of context for the repositories we maintain, see this list.

Content and wording of this page are open to discussion. happy to get suggestions especially from people who know more about these repos :)

Rendered

rustbot · 2025-07-17T11:27:58Z

r? @lcnr

rustbot has assigned @lcnr.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

apiraino · 2025-07-17T11:32:55Z

src/compiler/repositories.md

+
+The obvious focus of the Compiler Team is the [rustc] compiler itself, but the team also maintains a few other repositories:
+
+- An experimental [C codegen backend][c-backend] for rustc, which lowers Rust MIR to C code and compiles it with a C compiler. Currently maintained by X and Y.


Discussion:

Currently maintained by X and Y.

do all these "Currently maintained by X and Y" make sense? Do we have specific people more or less de-facto official maintainers for these repositories?

or just remove this wording?

It's an interesting question, because compiler team might have maintain-level access to a r-l/* repo, but doesn't necessarily actively maintain it as a whole team.

bjorn3 · 2025-07-17T11:33:01Z

src/compiler/repositories.md

+The obvious focus of the Compiler Team is the [rustc] compiler itself, but the team also maintains a few other repositories:
+
+- An experimental [C codegen backend][c-backend] for rustc, which lowers Rust MIR to C code and compiles it with a C compiler. Currently maintained by X and Y.
+- A [Cranelift][cranelift-backend] experimental rustc backend, based on [Cranelift]. This has the potential to improve compilation times in debug mode. Currently maintained by X and Y.


Effectively just me right now.

bjorn3 · 2025-07-17T11:33:32Z

src/compiler/repositories.md

+- A Rust port of the [LLVM::APFloat][rustc_apfloat] library. Maintained by X abnd Y.
+- A fork of [Enzyme], a high-performance automatic differentiator of LLVM and MLIR (more info at this [link][enzyme-mit]). The fork is currently maintained by X and Y.
+- [Thorin], a DWARF packaging utility supporting GNU extension and DWARF 5 package formats. Maintained by X and Y.
+- The [Rust Forge][forge], the documentation website you're reading. Maintained by the Rust project.


It seems like the GCC backend is missing from this list.

cg_gcc is probably not maintained by the compiler team as a whole, it's maintained by the
wg-gcc-backend WG

We as T-compiler have "only" write permissions on that repo (see list). Do you think this should be changed? Or maybe just add the repo to the list even if we don't have maintain role?

@antoyo and @GuillaumeGomez what do you think?

I'll let @antoyo confirm but on my end I think it's fine adding cg_gcc to the list.

We as T-compiler have "only" write permissions on that repo (see list). Do you think this should be changed? Or maybe just add the repo to the list even if we don't have maintain role?

I don't think compiler needs maintain access. Infra-wise, principle of least priviledges.

We as T-compiler have "only" write permissions on that repo (see list). Do you think this should be changed? Or maybe just add the repo to the list even if we don't have maintain role?

Note that the "maintain" role shouldn't ideally be used for anything, it is not necessary for most use-cases. So having the "maintain" role doesn't mean that a team maintains a specific repository. In fact even "write" doesn't mean that; maintaining something and having write access to it is a bit different, conceptually 😅

I'm OK with adding this to the list.

Note that the "maintain" role shouldn't ideally be used for anything, it is not necessary for most use-cases.

Interesting. Do you think then we should downgrade the role of T-compiler from maintain to write to make everything consistent?

In general, yes. There are some repos where maintain was needed for some specific permissions, but usually it shouldn't be required. It's hard to tell where downgrading is "safe" and where not though.

jieyouxu · 2025-07-17T11:30:50Z

src/compiler/repositories.md

+
+The obvious focus of the Compiler Team is the [rustc] compiler itself, but the team also maintains a few other repositories:
+
+- An experimental [C codegen backend][c-backend] for rustc, which lowers Rust MIR to C code and compiles it with a C compiler. Currently maintained by X and Y.


Remark: I think this is mostly a GSoC project or an experimental backend, I don't think compiler (as in the whole team) really maintains it. It's maintained by the codegen_c_maintainers subteam.

jieyouxu · 2025-07-17T11:35:05Z

src/compiler/repositories.md

+- A Rust port of the [LLVM::APFloat][rustc_apfloat] library. Maintained by X abnd Y.
+- A fork of [Enzyme], a high-performance automatic differentiator of LLVM and MLIR (more info at this [link][enzyme-mit]). The fork is currently maintained by X and Y.
+- [Thorin], a DWARF packaging utility supporting GNU extension and DWARF 5 package formats. Maintained by X and Y.
+- The [Rust Forge][forge], the documentation website you're reading. Maintained by the Rust project.


cg_gcc is probably not maintained by the compiler team as a whole, it's maintained by the
wg-gcc-backend WG

jieyouxu · 2025-07-17T11:37:37Z

src/compiler/repositories.md

+
+- An experimental [C codegen backend][c-backend] for rustc, which lowers Rust MIR to C code and compiles it with a C compiler. Currently maintained by X and Y.
+- A [Cranelift][cranelift-backend] experimental rustc backend, based on [Cranelift]. This has the potential to improve compilation times in debug mode. Currently maintained by X and Y.
+- A Rust port of the [LLVM::APFloat][rustc_apfloat] library. Maintained by X abnd Y.


Remark: I'm not sure about rustc_apfloat, it has licensing differences as it uses Apache 2 with LLVM exception. See https://github.com/rust-lang/rustc_apfloat#licensing.

jieyouxu · 2025-07-17T11:38:17Z

src/compiler/repositories.md

+- An experimental [C codegen backend][c-backend] for rustc, which lowers Rust MIR to C code and compiles it with a C compiler. Currently maintained by X and Y.
+- A [Cranelift][cranelift-backend] experimental rustc backend, based on [Cranelift]. This has the potential to improve compilation times in debug mode. Currently maintained by X and Y.
+- A Rust port of the [LLVM::APFloat][rustc_apfloat] library. Maintained by X abnd Y.
+- A fork of [Enzyme], a high-performance automatic differentiator of LLVM and MLIR (more info at this [link][enzyme-mit]). The fork is currently maintained by X and Y.


Remark: Enzyme is mostly maintained by @ZuseZ4, I don't think it's really maintained by the compiler team as a whole.

Also "maintain" is a strong word, I update our fork every once in a while when I see relevant improvements in Enzyme. I just have two extra commits that I carry along till I find a better location to push them to.
But I assume most of our forks are like that?

lcnr · 2025-07-17T12:08:55Z

r? @jieyouxu

Document repositories maintained by the T-compiler

2c84e6e

rustbot assigned lcnr Jul 17, 2025

rustbot added the S-waiting-on-review Status: Awaiting review from the assignee but also interested parties. label Jul 17, 2025

apiraino commented Jul 17, 2025

View reviewed changes

bjorn3 reviewed Jul 17, 2025

View reviewed changes

jieyouxu reviewed Jul 17, 2025

View reviewed changes

rustbot assigned jieyouxu and unassigned lcnr Jul 17, 2025


		The obvious focus of the Compiler Team is the [rustc] compiler itself, but the team also maintains a few other repositories:

		- An experimental [C codegen backend][c-backend] for rustc, which lowers Rust MIR to C code and compiles it with a C compiler. Currently maintained by X and Y.

Document repositories maintained by the T-compiler #918

Are you sure you want to change the base?

Document repositories maintained by the T-compiler #918

Conversation

apiraino commented Jul 17, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rustbot commented Jul 17, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jieyouxu Jul 17, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

apiraino Jul 17, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Kobzol Jul 17, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Kobzol Jul 17, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

lcnr commented Jul 17, 2025

Uh oh!

Uh oh!

apiraino commented Jul 17, 2025 •

edited

Loading

jieyouxu Jul 17, 2025 •

edited

Loading

apiraino Jul 17, 2025 •

edited

Loading

Kobzol Jul 17, 2025 •

edited

Loading

Kobzol Jul 17, 2025 •

edited

Loading