A workspace member with several crate types fails builds of unrelated packages
Maintainers usually reply within 1 day
Nobody has claimed this yet.
Assessment
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Newbie friendliness
- 58/100
- Issue type
- Bug
- Clarity
- Mostly clear
- Activity status
- Active
- Tech stack
- rust
- Domain
- build-system
Research direction
Start by locating the client-side workspace crate-type check: the report says it runs before the daemon starts and appears to inspect every workspace member. Trace how cargo rustd build -p app selects packages, then compare that with the crate-type validation. Done means the unrelated ffi member no longer blocks app; the report also describes the expected behavior when ffi is a dependency.
Written by the indexing model from the issue text.
Description
Summary
cargo rustd build -p <crate> fails for the whole workspace when any workspace member declares more than one crate type, for example crate-type = ["lib", "cdylib", "staticlib"]. The failing member does not need to be a dependency of the crate being built.
This pattern is common for crates that expose an FFI surface (UniFFI, cbindgen, JNI) and also keep a lib target for Rust consumers.
Reproduction
Environment: macOS arm64 (Darwin), rustd from the nightly release, Rust 1.99.0 with rust-src.
Create a workspace with two members. app is a plain binary. ffi is a library with three crate types. app does not depend on ffi.
Cargo.toml
[workspace]
members = ["app", "ffi"]
resolver = "3"
ffi/Cargo.toml
[package]
name = "ffi"
version = "0.1.0"
edition = "2024"
[lib]
crate-type = ["lib", "cdylib", "staticlib"]
ffi/src/lib.rs
pub fn add(a: u32, b: u32) -> u32 {
a + b
}
app/Cargo.toml
[package]
name = "app"
version = "0.1.0"
edition = "2024"
app/src/main.rs
fn main() {
println!("{}", 1 + 2);
}
Then run:
cargo +1.99.0 rustd build -p app
Actual result
error: `ffi` is built as `lib` and `cdylib` and `staticlib`, which the daemon does not link yet: it builds a `lib`, an `rlib` or a `bin` alone
Expected result
Either of these would help:
- Building
appsucceeds, becauseffiis not in its dependency graph. - If
ffiis a dependency,rustdbuilds itslibtarget and ignores thecdylibandstaticliboutputs when no requested target needs them.
Notes from isolating it
| Variant | Result |
|---|---|
Both crates are workspace members, ffi has three crate types |
Fails, error above |
ffi removed from members (so it is not in the workspace) |
cargo rustd build -p app succeeds |
Both are members, ffi has the default crate type |
cargo rustd build -p app succeeds |
So the check seems to run over every workspace member, not only the packages selected by -p or the dependencies of those packages. I have not read the source, so this is an inference from the behavior.
I also saw that the error is printed even when the daemon is not running (cargo rustd stop reports daemon: not running), so the check looks like it happens in the client before the daemon starts. This is from observation only.
Why it matters
The README already lists "workspace crates built as cdylib, dylib or staticlib" under "What is not supported", so I understand this is a known limit. The part that may be unintended is its reach. A workspace that ships an FFI crate next to other packages cannot use rustd at all today, even for packages that do not touch that crate. The README does not say that the check applies to the whole workspace.
Related: #1 (macOS support). The trivial workspace above builds and runs under rustd on macOS arm64 once ffi is excluded, which is useful data for that issue.
Happy to test a fix or a branch.
- Dominant language
- Rust
- Stars
- 248
- Forks
- 13
- Avg merge
- 7h 52m
- Merged PRs (30d)
- 5
Getting set up
- No Dockerfile or Docker Compose file
- No pull request template
- Read the contributing guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from DioxusLabs/rustd
-
Difficulty 1/5 Under an hour Newbie friendliness 78/100
DioxusLabs/rustd#15 · 1 comment ·
Maintainers usually reply within 1 day
-
Difficulty 4/5 3-5 days Newbie friendliness 48/100
DioxusLabs/rustd#28 ·
Maintainers usually reply within 1 day
-
Difficulty 4/5 3-5 days Newbie friendliness 35/100
DioxusLabs/rustd#27 ·
Maintainers usually reply within 1 day
-
Workspace fails to load: rquickjs types give layout of …: HasErrorType, and rustd never falls backOpen
Difficulty 4/5 3-5 days Newbie friendliness 25/100
DioxusLabs/rustd#26 ·
Maintainers usually reply within 1 day
-
Difficulty 3/5 1-2 days Newbie friendliness 65/100
DioxusLabs/rustd#25 ·
Maintainers usually reply within 1 day
All issues in DioxusLabs/rustd
Similar issues
-
bug rust
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
Maintainers usually reply within 2 days
-
Difficulty 2/5 1-3 hours Newbie friendliness 74/100
oxc-project/oxc-htmlish-parser#34 ·
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
longbridge/nsh#2541 ·
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 73/100
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 88/100