Refactor Dockerfile, test scripts and CI (WIP) #1101
Conversation
| @@ -1,15 +1,11 @@ | |||
| .git | |||
| * | |||
| openssl aes-256-cbc -K $encrypted_55750ae1fbc7_key -iv $encrypted_55750ae1fbc7_iv -in .github_deploy_key.enc -out $SSH_KEY_FILE -d; | ||
| chmod 600 $SSH_KEY_FILE; | ||
| printf "%s\n" \ | ||
| "Host github.com" \ |
There was a problem hiding this comment.
Personally, I would be a fan of moving these long multi-line bash scripts to a separate file, and invoking them. Keeps the yaml more readable.
There was a problem hiding this comment.
I don't mind having the details in here. I like the transparency of a single file.
|
|
||
| FROM base as testing | ||
| ENV MONGO_MAJOR=3.2 \ | ||
| MONGO_VERSION=3.2.9 |
| # /etc/hosts is corrupted if it has lines starting with tab. | ||
| # Exit to allow docker to restart. | ||
| if grep -P "^\t" /etc/hosts; then | ||
| echo "Host mapping in /etc/hosts is buggy, fail contain start." |
| @@ -9,4 +9,5 @@ pytest-mock==1.6.0 | |||
| pytest-watch==3.8.0 | |||
| pytest==2.8.5 | |||
| requests_mock==1.3.0 | |||
| six>=1.11.0 | |||
There was a problem hiding this comment.
I'm guessing it's here because the GCS storage plugin uses it.
There was a problem hiding this comment.
Maybe we can throw a quick comment - lines beginning with # are ignored
|
@gsfr @ambrussimon Let's please not monkey with the docker folder on this pass. There's enough complexity as-is. I'll be creating corresponding branches in downstream integration projects now. |
|
@ambrussimon |
| - tests/bin/setup-integration-tests-ubuntu.sh | ||
| - DOCKER_DIR="$HOME/.cache/docker" | ||
| - secure: HhT1TdJcpqys8juVMw/DIZeK7oD4595TEKH5KlowH7MvwwFAUyQFb5W63F8dgk7elvRG+3fmga/m1JfXO+Iu7PVD912eiNDagW9aB3CEl3Z8zg+JUL8IjpMCkyKQDyJMnfOkrzdxdaqfOK+WmF+13f2qBu9Kc7wdXuzgHQrg4+0= # CI_REGISTRY_USER | ||
| - secure: hh7VDZnkxgl/vqHtS4IpXfIAckKpVQvoCzNW7fstr5Mcu8KNiCWIPgObBRm+m13aqpcFTMWQ6lT2kzORz2wWRbDeVhI1eGWOJswGNHPHZLO0Jaei6yfY2nY2mpxZbl+vdg00jkN64mi1ab3e++QgeLFruW0gyNefXX7E5L/mHTs= # CI_REGISTRY_PASS |
There was a problem hiding this comment.
Should we replace these encrypted variables with Travis env variables, as used in the openssl line below?
| - if [ "$TRAVIS_EVENT_TYPE" == "push" -a "$TRAVIS_BRANCH" == "master" ]; then | ||
| docker tag core:dist scitran/core:latest; | ||
| docker push scitran/core:latest; | ||
| fi |
There was a problem hiding this comment.
Is there a good way to further clean up the conditionals in the after_success block? It's currently rather hard to read, but I do realize there are order dependencies that add complexity. At minimum, let's have consistency in the test order of $TRAVIS_BRANCH and $TRAVIS_EVENT_TYPE.
Current progress of backports.
Review Checklist