There should be no reason to use gopkg.in versioned imports now that
we're using go modules. I think, IANAE.
gopkg.in kind of gets in the way of modules, as it only pulls over
tagged releases from github.com -- this then means that you need to use
go modules 'replace' syntax in the go.mod to use a non-versioned commit
or branch. This is feasible, but kind of ugly.
go modules defaults to pulling the latest version, so the default
behavior is the same as when pulling go-jira.v1 from gopkg.in.
This README is huge; the document shouldn't take more than two or three
screens. If we need more docs, they should be linked to and live
elsewhere.
For now, start by removing unnecessary sections. Gitter has been
completely unused since 2017, so remove it. The table of contents also
takes a lot of vertical space, when we don't want to have that much
content to begin with.
Tips on forking are also unnecessary, since that's the same for all Go
projects. Moreover, this is out of date now that the project is a Go
module.
Remove the help output, as that can be obtained extremely easily by just
running 'jira -h'.
Finally, remove the v0-v1 changes, as v0 was last released over three
years ago. It's extremely unlikely that anyone still has to upgrade from
the older version.
And remove vendor/, as it's now unnecessary. go.sum ensures that
dependencies aren't tampered with, and proxy.golang.org keeps copies of
all the archives.
The following two files contain characters which aren't valid in source
files within a Go module:
t/.password-store/GoJira/api-token:gojira@corybennett.org.gpg
t/.password-store/GoJira/api-token:mothra@corybennett.org.gpg
The directory only contains test scripts, so it doesn't need to be
included in the module. The simplest way to do that is to start the
directory with an underscore.
Fixes#228.
We can also use the apt addon to install packages. We also don't need
fast_finish, since we don't use allow_failures anywhere.
Finally, the 'go get' line was pointless, as all dependencies are
vendored, and 'go test' will catch build failures.
Patrick Pichler