Hello,
I’m trying to speed up our build, and I’m trying to understand why Travis is duplicating a build in our matrix configuration. The first build specified in the matrix is getting duplicated.
language: ruby
cache: bundler
sudo: required
rvm:
- 2.2.4
services:
- docker
if: branch = master OR type = pull_request
env:
global:
- TZ=US/Eastern
- RAILS_ENV=test
- CC_TEST_REPORTER_ID=<redacted>
- CC_DOWNLOAD_URL=https://codeclimate.com/downloads/test-reporter/test-reporter-latest-linux-amd64
- CI=true
- TEST_RUNNERS=5
- RUBY_GC_HEAP_INIT_SLOTS=1800000
- RUBY_HEAP_FREE_MIN=18000
- RUBY_HEAP_SLOTS_INCREMENT=1000000
- RUBY_HEAP_SLOTS_GROWTH_FACTOR=1
- RUBY_GC_MALLOC_LIMIT=1000000000
matrix:
- TEST_SUITE=test:units0 test:agents
- TEST_SUITE=test:units1 test:workers
- TEST_SUITE=test:units2
- TEST_SUITE=test:units3 test:functionals
- TEST_SUITE=test:units4
script: bundle exec rake $TEST_SUITE
before_install:
- curl -L ${CC_DOWNLOAD_URL} > ./cc-test-reporter
- chmod +x ./cc-test-reporter
- ./cc-test-reporter before-build
- sudo apt-get install libaspell-dev libicu-dev
- gem install bundler -v 1.17.3
- git clone --depth 1 git@github.com:customink/oracle-scripts.git
- source ./oracle-scripts/travis_install.sh
before_script:
- mkdir -p tmp/cache
- ./oracle-scripts/wait_for_oracle.sh
- ./bin/setup
after_script:
- if [[ "$TRAVIS_TEST_RESULT" == 0 ]]; then ./cc-test-reporter sum-coverage coverage/codeclimate.*.json fi
- if [[ "$TRAVIS_TEST_RESULT" == 0 ]]; then ./cc-test-reporter sum-coverage --output - --parts 6 /tmp/codeclimate-test-coverage*.json | ./cc-test-reporter upload-coverage --input -; fi
- ./cc-test-reporter after-build -r $CODECLIMATE_REPO_TOKEN --exit-code $TRAVIS_TEST_RESULT
stages:
- prepare cache
- test
jobs:
include:
- stage: prepare cache
script: true
before_script: skip
after_script: skip
- stage: test
I’m trying to speed up our build and get code-coverage working in this setup, but the documentation is not helping me whatosever in getting the code-climate coverage report configured, or why I have a duplicate build.