- When reporting issues, please include the output of
atlantis version
. - Also include the steps required to reproduce the problem if possible and applicable. This information will help us review and fix your issue faster.
- When sending lengthy log-files, consider posting them as a gist (https://gist.github.com). Don't forget to remove sensitive data from your logfiles before posting (you can replace those parts with "REDACTED").
We take security issues seriously. Please email us directly at security [at] runatlantis.io instead of opening an issue.
Get the source code:
go get github.com/runatlantis/atlantis
This will clone Atlantis into $GOPATH/src/github.com/runatlantis/atlantis
(where $GOPATH
defaults to ~/go
).
Go to that directory:
cd $GOPATH/src/github.com/runatlantis/atlantis
Compile Atlantis:
go install
Run Atlantis:
atlantis server --gh-user <your username> --gh-token <your token> --repo-whitelist <your repo> --gh-webhook-secret <your webhook secret> --log-level debug
If you get an error like command not found: atlantis
, ensure that $GOPATH/bin
is in your $PATH
.
- Create a test terraform repository in your GitHub.
- Create a personal access token for Atlantis. See Create a GitHub token.
- Start Atlantis in server mode using that token:
atlantis server --gh-user <your username> --gh-token <your token> --repo-whitelist <your repo> --gh-webhook-secret <your webhook secret> --log-level debug
- Download ngrok from https://ngrok.com/download. This will enable you to expose Atlantis running on your laptop to the internet so GitHub can call it.
- When you've downloaded and extracted ngrok, run it on port
4141
:
ngrok http 4141
- Create a WebHook in your repo and use the
https
url thatngrok
printed out after runningngrok http 4141
. Be sure to append/events
so your webhook url looks something likehttps://efce3bcd.ngrok.io/events
. See Add GitHub Webhook. - Create a pull request and type
atlantis help
. You should see the request in thengrok
and Atlantis logs and you should also see Atlantis comment back.
ctx.Log
should be available in most methods. If not, pass it down.- levels:
- debug is for developers of atlantis
- info is for users (expected that people run on info level)
- warn is for something that might be a problem but we're not sure
- error is for something that's definitely a problem
- ALWAYS logs should be all lowercase (when printed, the first letter of each line will be automatically capitalized)
- ALWAYS quote any string variables using %q in the fmt string, ex.
ctx.Log.Info("cleaning clone dir %q", dir)
=>Cleaning clone directory "/tmp/atlantis/lkysow/atlantis-terraform-test/3"
- NEVER use colons "
:
" in a log since that's used to separate error descriptions and causes- if you need to have a break in your log, either use
-
or,
ex.failed to clean directory, continuing regardless
- if you need to have a break in your log, either use
- ALWAYS use lowercase unless the word requires it
- ALWAYS use
errors.Wrap(err, "additional context...")"
instead offmt.Errorf("additional context: %s", err)
because it is less likely to result in mistakes and gives us the ability to trace call stacks - NEVER use the words "error occurred when...", or "failed to..." or "unable to...", etc. Instead, describe what was occurring at time of the error, ex. "cloning repository", "creating AWS session". This will prevent errors from looking like
Error setting up workspace: failed to run git clone: could find git
and will instead look like
Error: setting up workspace: running git clone: no executable "git"
This is easier to read and more consistent
- place tests under
{package under test}_test
to enforce testing the external interfaces - if you need to test internally i.e. access non-exported stuff, call the file
{file under test}_internal_test.go
- use our testing utility for easier-to-read assertions:
import . "github.com/runatlantis/atlantis/testing"
and then useAssert()
,Equals()
andOk()
- don't try to describe the whole test by its function name. Instead use
t.Log
statements:
// don't do this
func TestLockingWhenThereIsAnExistingLockForNewEnv(t *testing.T) {
...
// do this
func TestLockingExisting(t *testing.T) {
t.Log("if there is an existing lock, lock should...")
...
t.Log("...succeed if the new project has a different path") {
// optionally wrap in a block so it's easier to read
}
- each test should have a
t.Log
that describes what the current state is and what should happen (like a behavioural test)
- Update version number in
main.go
- Update
CHANGELOG.md
with latest release number and information (this URL might be useful: https://github.com/runatlantis/atlantis/compare/v0.3.5...master) - Create a pull request and merge to master
- Check out master and fetch latest
- Run
make release
- Go to https://github.com/runatlantis/atlantis/releases and click "Draft a new release"
- Prefix version with
v
- The title of the release is the same as the tag (ex. v0.2.2)
- Fill in description by copying from the CHANGELOG just without the Downloads section
- Drag in binaries made with
make release
- Prefix version with
- Re-run master branch build to ensure tag gets pushed to Docker hub: https://hub.docker.com/r/runatlantis/atlantis/tags/