Crates.io | zmq-pw |
lib.rs | zmq-pw |
version | 0.9.8 |
source | src |
created_at | 2017-06-22 17:05:12.200576 |
updated_at | 2017-08-02 10:15:54.079847 |
description | High-level bindings to the zeromq library |
homepage | |
repository | https://github.com/evernym/rust-zmq-pw |
max_upload_size | |
id | 20185 |
size | 169,050 |
Rust ZeroMQ bindings.
Currently, rust-zmq requires ZeroMQ 3.2 or newer. For example, on recent Debian-based distributions, you can use the following command to get the prerequiste headers and library installed:
apt install libzmq3-dev
If your OS of choice does not provide packages of a new-enough libzmq, you will first have to install it from source; see https://github.com/zeromq/libzmq/releases.
rust-zmq uses cargo to install. Users should add this to
their Cargo.toml
file:
[dependencies]
zmq = "0.8"
The build normally uses pkg-config
to find out about libzmq's
location. If that is not available, the environment variable
LIBZMQ_PREFIX
(or alternatively, LIBZMQ_LIB_DIR
and
LIBZMQ_INCLUDE_DIR
) can be defined to avoid the invocation of
pkg-config
.
rust-zmq
is a pretty straight forward port of the C API into Rust:
extern crate zmq;
fn main() {
let ctx = zmq::Context::new();
let mut socket = ctx.socket(zmq::REQ).unwrap();
socket.connect("tcp://127.0.0.1:1234").unwrap();
socket.send_str("hello world!", 0).unwrap();
}
You can find more usage examples in https://github.com/erickt/rust-zmq/tree/master/examples.
Install for contributing to rust-zmq:
% git clone https://github.com/erickt/rust-zmq
% cd rust-zmq
% cargo build
Note that the master
branch is currently in API-breaking mode while
we try to make the API more ergomic and flexible for the 0.9
release
series.
This means that pull requests (e.g. bugfixes), which do not need to
break API should be submitted for the release/v0.8
branch. This
also applies to new features, if they can be implemented in an
API-compatible way, the pull request should also aim for
release/v0.8
. Please submit an issue for missing features before you
start coding, so the suitable branch and other potential questions can
be clarified up-front.
The reason for using branches, and thus increasing overhead a bit for
all involved, is that it's not yet clear how long it will take to
reach a point in master
that we feel comfortable with releasing as
0.9.0, as we'd like to have the core part of the API more-or-less
fixed by then. Using the release/v0.8
branch we can deliver bugfixes
and smaller features in the meantime without forcing users to follow
master's changing API and to continuously adjust their code to API
changes.