289b30715d
This commit converts the dependency management from Godeps to the vendor folder, also switching the tool from godep to trash. Since the upstream tool lacks a few features proposed via a few PRs, until those PRs are merged in (if), use github.com/karalabe/trash. You can update dependencies via trash --update. All dependencies have been updated to their latest version. Parts of the build system are reworked to drop old notions of Godeps and invocation of the go vet command so that it doesn't run against the vendor folder, as that will just blow up during vetting. The conversion drops OpenCL (and hence GPU mining support) from ethash and our codebase. The short reasoning is that there's noone to maintain and having opencl libs in our deps messes up builds as go install ./... tries to build them, failing with unsatisfied link errors for the C OpenCL deps. golang.org/x/net/context is not vendored in. We expect it to be fetched by the user (i.e. using go get). To keep ci.go builds reproducible the package is "vendored" in build/_vendor. |
||
|---|---|---|
| .. | ||
| LICENSE.md | ||
| README.md | ||
| wordwrap.go | ||
go-wordwrap
go-wordwrap (Golang package: wordwrap) is a package for Go that
automatically wraps words into multiple lines. The primary use case for this
is in formatting CLI output, but of course word wrapping is a generally useful
thing to do.
Installation and Usage
Install using go get github.com/mitchellh/go-wordwrap.
Full documentation is available at http://godoc.org/github.com/mitchellh/go-wordwrap
Below is an example of its usage ignoring errors:
wrapped := wordwrap.WrapString("foo bar baz", 3)
fmt.Println(wrapped)
Would output:
foo
bar
baz
Word Wrap Algorithm
This library doesn't use any clever algorithm for word wrapping. The wrapping is actually very naive: whenever there is whitespace or an explicit linebreak. The goal of this library is for word wrapping CLI output, so the input is typically pretty well controlled human language. Because of this, the naive approach typically works just fine.
In the future, we'd like to make the algorithm more advanced. We would do so without breaking the API.