build: fix: configure fails if GOPATH environment variable not set

If the GOPATH enviroment variable was not set, go uses the default
GOPATH (~/go/).

The configure script was only checking if the GOPATH environment
is set. If it wasn't the script was failing.

Instead of checking if the GOPATH environment variable is set, check if
"go env GOPATH" returns a non-emtpy string.

* Access token forwarding through nginx auth request

Related to #420.

(cherry picked from commit b138872beaaa7f47d43a1c0fef11a67f57e61eff)
Signed-off-by: David Holsgrove <david.holsgrove@biarri.com>

* Improved documentation for auth request token

(cherry picked from commit 6fab314f7203f4d652bb34247abb4e7cb497c41d)
Signed-off-by: David Holsgrove <david.holsgrove@biarri.com>

* Update README.md

Example should set header as `X-Access-Token`

Co-Authored-By: davidholsgrove <davidholsgrove@users.noreply.github.com>

* Update Changelog to reference https://github.com/pusher/oauth2_proxy/pull/68

* Fix Changelog message location

Add oidc-issuer-url arg to README

Update release notes for v3.1.0

Add Quay links to ARM repositories

Ensure binary archives are gzipped

feat(arm): Cross build arm and arm64 docker images

Change cookie index separator to underscore

Whitelist domains

Co-Authored-By: marratj <marrat@marrat.de>

- Requires `qemu-user-static`, added to travis - maybe incorrect?
- Add build guide
- `.gitignore` `release/` directory

feat(arm): Makefile add armv6 and arm64 to releases

Returns HTTP unauthorized for ajax requests instead of redirecting to the sign-in page

* Implemented flushing interval

When proxying streaming responses, it would not flush the response writer buffer until some seemingly random point (maybe the number of bytes?). This makes it flush every 1 second by default, but with a configurable interval.

* flushing CHANGELOG

* gofmt and goimports

watcher: properly break out in nested blocks

Found via staticcheck:

watcher.go:48:5: ineffective break statement. Did you mean to break out of the outer loop? (SA4011)

* redirect to original path after login

* tests for new redirect behaviour

* fixed comment

* added redirect fix to changelog