I often want to compile in release mode with debug = true so that I can read the generated assembly a bit easier. I am currently doing this:
[profile.release]
debug = true
but I don't want any debug symbols in my final release build. I'd like to do something like:
[profile.custom]
debug = true
opt-level = 3
rpath = false
lto = true
debug-assertions = false
codegen-units = 1
panic = 'unwind'
And then run
cargo build --custom
I've read the documentation to no avail.
As of Rust v1.57.0, the custom profiles feature is now stable.
Add a profile section, specify a base profile to inherit from, and tweak as you see fit:
[profile.production]
inherits = "release"
lto = true
Specify the profile to use via the --profile <name> flag to cargo.
Does Cargo support custom profiles?
No, stable releases of Cargo do not support this. It is available as an unstable nightly feature.
If you are using a nightly version of Cargo, you can create custom profiles in your Cargo.toml:
cargo-features = ["named-profiles"]
[profile.release-lto]
inherits = "release"
lto = true
And then use them:
cargo +nightly build --profile release-lto -Z unstable-options
Related
I have just started looking into the Bevy game engine for Rust. It has a feature called dynamic, which enables dynamic linking, to speed up compilation time during development. We are, however, advised to disable this when building for release.
Is there a way to tell Cargo to enable the dynamic feature for a debug build, but disable it for a release build? Or do I have to personally remember to change bevy = { version = "0.5.0", features = ["dynamic"] } to bevy = "0.5.0" in Cargo.toml before running cargo build --release?
Along the lines of Rodrigo's comment, can confirm the following seems to work well:
[dependencies]
bevy = { version = "0.5.0" }
[features]
default = ["fast-compile"]
fast-compile = ["bevy/dynamic"]
Then for development, simply: cargo build
And for release: cargo build --release --no-default-features
Cargo.toml:
[features]
parallel = ["rayon"]
[dependencies.rayon]
version = "1.5"
optional = true
lib.rs:
#[cfg(feature = "parallel")]
pub mod par;
Rust Analyzer:
code is inactive due to #[cfg] directives: feature = "parallel" is disabled
How to enable the optional dependency?
You can set the Rust Analyzer configuration option rust-analyzer.cargo.features to an array containing a list of features that you want RA to consider active.
You can also set rust-analyzer.cargo.allFeatures to true, in order to enable all features in the project.
Methods for setting these vary according to the IDE you are using - for example, if using VS Code, you can set it via the "Extension Settings" for Rust Analyzer.
This assumes you are asking how to activate the features within Rust Analyzer - to activate them when building or running from Cargo, just use the --features option.
See: Rust Analyzer Manual
All non-default features need to be specified when running. So run with:
# Option 1
cargo run --features parallel
# Option 2
cargo run --all-features
Check cargo run --help for more help.
I want to compile two binaries for different target architectures (eg. Skylake and Sandy Bridge).
These are usually two lengthy cargo commands:
RUSTFLAGS="-C target-cpu=skylake" cargo build --target
x86_64-unknown-linux-gnu --release
How can I set up cargo to build both binaries (with different names) from the same main.rs automatically?
Ideally in either the config.toml or the Cargo.toml so I can add it to a repository.
You can add the following text to your config.toml:
[build]
target = x86_64-unknown-linux-gnu
rustflags = ["-C","target-cpu=skylake"]
[profile.dev] #do not need to add `--release` now
opt-level = 3
debug = false
debug-assertions = false
overflow-checks = false
lto = false
panic = 'unwind'
incremental = false
codegen-units = 16
rpath = false
But it seems like it can't compile for two different target architectures with one config.toml, so you may have to create two config.toml and use cargo --manifest-path PATH/TO/CONFIG to compile two binaries separately.
The program I'm writing runs much faster when the -C target-cpu=native flag is passed to rustc. I want to give users a simple, platform-independent way to enable this when compiling, so I added a Cargo feature cpu_native = [] in Cargo.toml and created this Cargo config in my project:
[target.'cfg(cpu_native)']
rustflags = ["-C", "target-cpu=native"]
However, this has no effect on my program, and passing --features cpu_native to Cargo does not even trigger a recompile. Changing to the following Cargo config does force re-compilation with faster instructions:
[build]
rustflags = ["-C", "target-cpu=native"]
However, this will compile with target-cpu=native with the default Cargo features, which was not what I wanted. From the Cargo book, what I want seems to be possible, but I don't see what I'm doing wrong.
I don't think this is supported (yet?). I enhanced Cargo to print out what config flags are checked against when resolving:
[
Name("debug_assertions"),
Name("proc_macro"),
KeyPair("target_arch", "x86_64"),
KeyPair("target_endian", "little"),
KeyPair("target_env", ""),
KeyPair("target_family", "unix"),
KeyPair("target_os", "macos"),
KeyPair("target_pointer_width", "64"),
Name("unix"),
]
[target.'cfg(cpu_native)']
This is the incorrect syntax for a Cargo feature; it would normally be cfg(feature = "cpu_native").
I want to conditionally compile my source code using cfg with Cargo,
after Googling for a while,
it seems that the solution is to use cargo --features.
http://doc.crates.io/manifest.html
I tried adding a few
#[cfg(feature = "foo")]
in the source code and
cargo build --features foo
, but it says
Package `xxx v0.0.1 (file:///C:/yyy/xxx)` does not have these features: `foo`
How can I let cargo identify the features? Do I have to add something in Cargo.toml?
Here's the version of rustc and cargo I am using:
C:\>rustc --version
rustc 0.13.0-nightly (42deaa5e4 2014-12-16 17:51:23 +0000)
C:\>cargo --version
cargo 0.0.1-pre-nightly (5af754d 2014-12-18 01:50:48 +0000)
You have to introduce the existing features in your Cargo.toml.
I was able to conditionally compile by doing the following:
In Cargo.toml, create a features section and introduce a certain feature name:
[features]
customfeature = [] # feature has no explicit dependencies
If you want your feature to have specific dependencies check the examples in the documentation.
In your code, use #[cfg(feature="customfeature")]
Run cargo build --features customfeature
Since your steps 2 & 3 seem to be fine, there must probably be a problem with your Cargo.toml.
As stated in other answers, you can use features for this. I would like to add that features do not only allow you to conditionally compile parts of your code but also to conditionally include dependencies that may be part of that code. Consider the following snippets:
You can activate the conditional code using a feature flag as already described in other anwsers:
cargo build --features customfeature
You need to mark your conditional code to exist only when your customfeature is enabled:
#[cfg(feature = "customfeature")]
fn my_func() {
my_optional_dependency::do_something();
}
// This includes dependencies only when customfeature is enabled
#[cfg(feature = "customfeature")]
extern crate my_optional_dependency;
....
#[cfg(feature = "customfeature")]
use my_optional_dependency::*;
....
Your Cargo.toml needs to have the following sections:
[dependencies.my_optional_dependency]
version = "1.2.3"
optional = true
[features]
customfeature = ["my_optional_dependency"]
This allows you to activate certain parts of your code along with their dependencies only if a feature is enabled.
Alternatively, you could create a cargo configuration file in your project, by creating a .cargo subdir in your project main folder, adding in it a config.toml file,
then inserting this section in .cargo/config.toml:
[build]
rustflags = "--cfg my_cfg_flag"
This will make cargo call rustc with flags --cfg my_cfg_flag
See here for details:
https://doc.rust-lang.org/cargo/reference/config.html