This post is an overview of the major projects the Cargo team is interested in tackling in 2020.
It can be difficult to plan and predict around a volunteer-based open-source project with limited resources. Instead of trying to present a wish list, these are projects that already have a solid effort planned to push them forward. That doesn't mean that we are not interested in other projects. We have compiled a more detailed wish list at https://github.com/rust-lang/cargo/projects/1 that gives an outline of things we would like to see, but are unlikely to have significant progress this year.
If you are interested in helping, please let us know! We may not have time to shepherd additional projects, but we may have time to give some amount of feedback and review, particularly for well-motivated people who can do the legwork of design and gathering a consensus.
Features provide a way to express optional dependencies and conditional compilation of code. Fixes and enhancements to Features are one of the most requested things we hear. In the beginning of 2020, we plan to implement a new feature resolver which will make it easier to make progress on implementing and experimenting with new behavior. There is a wide variety of different enhancements that we are looking at, which we hope to make incremental progress on while retaining a full picture of the long-term plan.
Initially we plan to address the issues of decoupling shared dependencies
built with different features. Currently, features are unified for all uses of
a dependency, even when it is not necessary. This causes problems when a
feature intended for one context is incompatible with another. This often
happens for packages which have conditional no_std support. This appears
with build-dependencies, dev-dependencies, target-specific dependencies, and
large workspaces, each of which have their unique challenges.
Beyond that, the following is a brief view of the other major enhancements we are tracking for the future:
There are some significant challenges around retaining backwards compatibility, and being sensitive to increased build times. We hope that we can address some of the major pain points while balancing those concerns.
The "std aware Cargo" project is to make Cargo aware of the Rust standard
library, and to build it from source instead of using the pre-built binaries
that ship with rustc. Some of the notable benefits are:
A significant amount of work has already been finished in 2019 with the
-Zbuild-std feature available on the nightly channel. There is still a
long road to bring it to a state where it can be stabilized. Work is being
tracked in the wg-cargo-std-aware repo, and anyone interested is
encouraged to leave feedback on the issue tracker.
Profiles have received a significant amount of work in 2018 and 2019. Overrides are now stable (shipping in Rust 1.41). Custom named profiles are available on the nightly channel. In 2020 we hope to continue pushing these enhancements forward. Some of the efforts we are working towards are:
build profile which can make it easier to define build-script settings.Some ongoing efforts don't have an end, and we intend to continue making progress with them. Several new chapters have been added to the documentation, and there is more to come. The JSON APIs are continually expanding with new information making it easier to integrate tools and extract information. And of course, trying to stay on top of bugs and issues!
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。