Pants v2 was designed with these use-cases in mind, and in particular with how they are utilized by engineering teams of various sizes, across different industries. In neither case will such a tool be able to address the needs of most real-world engineering teams today. The designs of existing tools were typically driven by one of two concerns: A) Support for small to medium codebases using a single language/framework, or B) Support for the internal needs of a single company. These include: a growing reliance on large tangles of third-party dependencies, more application frameworks and layers of abstraction, novel deployment paradigms (containers, cloud functions, microservices), and the emergence of data science as an engineering discipline in its own right.Įxisting tools were not designed with these trends in mind, and they are often limited in capability and scalability in the face of these and other innovations. Why do you think there is a need for another build system? What makes Pants unique in comparison to other solutions?īenjy Weinberger: In the last 5-10 years, a few convergent trends have made codebases and their build workflows increasingly complex. InfoQ: The build system arena is quite crowded. InfoQ has spoken with Benjy Weinberger, one of the creators of Pants alongside John Sirois, and currently CEO of Toolchain, Pants' main sponsor. Pants is mostly written in Python, with an ongoing Rust migration aimed to reduce client latency, and provides an API to enable the creation of plugins to add custom build steps, in case required. In addition to creating libraries and executables, Pants is also able to deploy to Docker and AWS Lambda using their corresponding backends.
".docformatter",Įach backend can be configured by providing specific options in its own.
For example, this is how you can specify a list of backends for a Python project, including linting and mypy support: This will tell Pants which source files to look for and create a suitable BUILD file for you. Indeed, thanks to a number of default choices, many projects will be ready to go by simply defining a pants.toml configuration file that includes the Pants version you want to pin your project to and required backends. One of Pants' goal is making build files less cluttered and light on the developer by avoiding unnecessary configuration.
This also makes it suitable for splitting work across multiple identically-configured hosts. Pants is compatible with servers complying with the caching and remote execution API, also used by Bazel and others. Pants breaks a build into many smaller tasks to be able to run them concurrently and cache their results to avoid having to run them again unless strictly required. Pants aims to be a fast, scalable, and user-friendly build system. In its upcoming release, now available to early adopters, build system Pants adds Java, Scala, and Go to previously supported Python.
0 kommentar(er)
