"go mod download" does not depend on the VERSION env var, so moving the ARG directive after the RUN will allow better use of the Docker build cache - subsequent builds on the same machine need only re-run the "go mod download" if go.mod or go.sum has changed, rather than re-running it any time the VERSION value passed from the Makefile has changed
Co-authored-by: Joel Speed <Joel.speed@hotmail.co.uk>
This includes a fix for our samesite cookie parsing. The behaviour
changed in 1.16 so that the default value now leaves it empty, so it's
equivalent to not setting it (as per spec)
* simplify github actions workflow
no more GOPATH, update Go to 1.15.x
* add script to install golangci-lint
* drop support for Go 1.14
* check docker build in ci
* update alpine linux to 3.12
* update CHANGELOG
* fix golangci-lint installation
Co-authored-by: Joel Speed <Joel.speed@hotmail.co.uk>
* Upgrade base image tags and golangci-lint
* Upgrade golang and golangci-lint in travis-ci
* fix: tests, pointed out by golangci-lint and format files
* Upgrade dependencies
* update changelog
* fix: tests related to https://github.com/pusher/oauth2_proxy/pull/418
* Separate tests using go version build tags
* Update CHANGELOG
* Revert "Separate tests using go version build tags"
This reverts commit 9b7e65eb90cae954dc7b6316345d3207205d488a.
* fix test to support go1.14 and go1.13
Co-authored-by: Joel Speed <Joel.speed@hotmail.co.uk>
* Add nsswitch.conf to Docker image
Created nsswitch.conf to use locally defined translations before DNS. Copied to /etc/nsswitch.conf in the image.
* Add new line
* Updated Changelog
Co-authored-by: Dan Bond <danbond@protonmail.com>
This will let Docker cache the results of the vendor dependencies.
Making re-builds during testing faster.
Also clean-up spurious test & rm in ./configure
Use simple USER directive.
Using `addgroup` in final `arm` image when building on amd64 doesn't work.
I must have made a mistake during cross build verification.
Alternative is to use qemu-static but it's not worth it for this.
Run as non-root user and group
In the unlikely event that you are currently persisting data to disk then this
change may break file read/write access due to a change in the UID/GID that the
oauth2_proxy process runs as.
Run as non-root system user and group `oauth2proxy` with UID/GID `2000` to avoid clashing with typical local users.
An alternative to creating a separate user is to ~~chown binary and~~ run as `USER nobody`, which also works, can amend this PR if required.
Least access privileges.
Close: https://github.com/pusher/oauth2_proxy/issues/78
Locally with Docker (`-version`):
```
$ ps aux | grep oauth2
2000 25192 6.0 0.0 0 0 ? Ds 15:53 0:00 [oauth2_proxy]
```
Running in Kubernetes 1.13 with the following also specified:
```
securityContext:
readOnlyRootFilesystem: true
runAsNonRoot: true
runAsUser: 10001
```
```
$ kubectl exec -it -n oauth2-proxy oauth2-proxy-85c9f58ffc-dz9lr sh
/opt $ whoami
whoami: unknown uid 10001
/opt $ ps aux
PID USER TIME COMMAND
1 10001 0:00 /opt/oauth2_proxy --whitelist-domain=.example.com --cookie-domain=example.com --email-domain=example.com --upstream=file:///dev/null --http-address=0.0.0.0:4180
11 10001 0:00 sh
17 10001 0:00 ps aux
```
<!--- Go over all the following points, and put an `x` in all the boxes that apply. -->
<!--- If you're unsure about any of these, don't hesitate to ask. We're here to help! -->
- [x] My change requires a change to the documentation or CHANGELOG.
- [x] I have updated the documentation/CHANGELOG accordingly.
- [x] I have created a feature (non-master) branch for my PR.
