discourse/.travis.yml

language: ruby

env:
  global:
    - DISCOURSE_HOSTNAME=www.example.com
    - RUBY_GC_MALLOC_LIMIT=50000000
  matrix:
    - "RAILS_MASTER=0"
    - "RAILS_MASTER=1"

addons:
  postgresql: 9.5
  apt:
    packages:
    - gifsicle
    - jpegoptim
    - optipng
    - jhead

matrix:
  allow_failures:
    - env: "RAILS_MASTER=1"
    - rvm: rbx-2
  fast_finish: true

rvm:
  - 2.3.1

services:
  - redis-server

sudo: required
dist: trusty

cache:
  directories:
    - vendor/bundle

before_install:
  - gem install bundler
  - git clone --depth=1 https://github.com/discourse/discourse-backup-uploads-to-s3.git plugins/discourse-backup-uploads-to-s3
  - git clone --depth=1 https://github.com/discourse/discourse-spoiler-alert.git plugins/discourse-spoiler-alert
  - git clone --depth=1 https://github.com/discourse/discourse-cakeday.git plugins/discourse-cakeday
  - git clone --depth=1 https://github.com/discourse/discourse-canned-replies.git plugins/discourse-canned-replies
  - npm i -g eslint babel-eslint
  - eslint app/assets/javascripts
  - eslint --ext .es6 app/assets/javascripts
  - eslint --ext .es6 test/javascripts
  - eslint --ext .es6 plugins/**/assets/javascripts
  - eslint test/javascripts

before_script:
  - bundle exec rake db:create db:migrate

install:
  - bash -c "if [ '$RAILS_MASTER' == '1' ]; then bundle update --retry=3 --jobs=3 arel rails seed-fu; fi"
  - bash -c "if [ '$RAILS_MASTER' == '0' ]; then bundle install --without development --deployment --retry=3 --jobs=3; fi"

script: "bundle exec rspec && bundle exec rake plugin:spec && bundle exec rake qunit:test['200000']"
Added TravisCI configuration file 2013-02-08 03:43:07 +08:00			`language: ruby`
some possible improvements to build times - use our new infrastructure based on LXC and EC2 - enable caching for bundler - use schema:load (best practice) - use --deployment (best practice when a Gemfile.lock is around) 2014-11-07 09:03:08 +08:00
Add RAILS_MASTER to the travis build matrix 2014-10-29 06:48:57 +08:00			`env:`
			`global:`
			`- DISCOURSE_HOSTNAME=www.example.com`
			`- RUBY_GC_MALLOC_LIMIT=50000000`
			`matrix:`
Run main tests on Travis first. 2015-07-17 16:23:43 +08:00			`- "RAILS_MASTER=0"`
Add RAILS_MASTER to the travis build matrix 2014-10-29 06:48:57 +08:00			`- "RAILS_MASTER=1"`

`NewPostManager` determines whether to queue a post or not 2015-04-01 00:58:56 +08:00			`addons:`
Use PostgreSQL 9.5 on travis. 2016-11-15 15:30:38 +08:00			`postgresql: 9.5`
Let's try @wil93's recommendation to whitelist packages 2015-08-14 16:52:33 +08:00			`apt:`
			`packages:`
			`- gifsicle`
			`- jpegoptim`
			`- optipng`
			`- jhead`
`NewPostManager` determines whether to queue a post or not 2015-04-01 00:58:56 +08:00
Add RAILS_MASTER to the travis build matrix 2014-10-29 06:48:57 +08:00			`matrix:`
			`allow_failures:`
Update travis config file A few things in here: - Test on ruby 2.2 - Allow failures in all variants of RAILS_MASTER test - Use db:create task 2015-04-25 04:28:27 +08:00			`- env: "RAILS_MASTER=1"`
Add Rubinius to Build Matrix with Allowed Failure Please add Rubinius to the build matrix with failure allowed. 2015-02-23 03:51:40 +08:00			`- rvm: rbx-2`
Add RAILS_MASTER to the travis build matrix 2014-10-29 06:48:57 +08:00			`fast_finish: true`

Added TravisCI configuration file 2013-02-08 03:43:07 +08:00			`rvm:`
Revert "Run Travis against 2.4.0 as well." This reverts commit 0000de9501d1e8e2a1a27028086e5631ece29905. 2017-01-09 12:21:53 +08:00			`- 2.3.1`
some possible improvements to build times - use our new infrastructure based on LXC and EC2 - enable caching for bundler - use schema:load (best practice) - use --deployment (best practice when a Gemfile.lock is around) 2014-11-07 09:03:08 +08:00
			`services:`
			`- redis-server`

get trusty working on travis 2016-05-23 10:04:44 +08:00			`sudo: required`
use trusty 2016-05-23 09:58:20 +08:00			`dist: trusty`
some possible improvements to build times - use our new infrastructure based on LXC and EC2 - enable caching for bundler - use schema:load (best practice) - use --deployment (best practice when a Gemfile.lock is around) 2014-11-07 09:03:08 +08:00
Lets see if this properly caches bundler on travis. 2015-08-17 01:05:57 +08:00			`cache:`
			`directories:`
			`- vendor/bundle`
some possible improvements to build times - use our new infrastructure based on LXC and EC2 - enable caching for bundler - use schema:load (best practice) - use --deployment (best practice when a Gemfile.lock is around) 2014-11-07 09:03:08 +08:00
Add JSHint to Travis validation Run before the install as it acts as a fast fail vs the gem install time 2014-03-08 14:14:29 +08:00			`before_install:`
Lets see if this properly caches bundler on travis. 2015-08-17 01:05:57 +08:00			`- gem install bundler`
Test important plugins on our travis build. 2016-08-18 19:15:39 +08:00			`- git clone --depth=1 https://github.com/discourse/discourse-backup-uploads-to-s3.git plugins/discourse-backup-uploads-to-s3`
Add more plugins to test for. 2016-08-19 16:54:33 +08:00			`- git clone --depth=1 https://github.com/discourse/discourse-spoiler-alert.git plugins/discourse-spoiler-alert`
Run tests for Discourse Cakeday on Travis. 2016-11-23 14:23:12 +08:00			`- git clone --depth=1 https://github.com/discourse/discourse-cakeday.git plugins/discourse-cakeday`
Run specs for canned replies as well. 2016-12-30 14:17:36 +08:00			`- git clone --depth=1 https://github.com/discourse/discourse-canned-replies.git plugins/discourse-canned-replies`
No need to pin eslint to 2.2 anymore. 2016-04-08 15:08:48 +08:00			`- npm i -g eslint babel-eslint`
Run eslint instead of jshint, remove rbx 2015-08-14 03:22:33 +08:00			`- eslint app/assets/javascripts`
			`- eslint --ext .es6 app/assets/javascripts`
			`- eslint --ext .es6 test/javascripts`
Start checking eslint in plugins. 2016-06-14 16:11:45 +08:00			`- eslint --ext .es6 plugins/**/assets/javascripts`
Run eslint instead of jshint, remove rbx 2015-08-14 03:22:33 +08:00			`- eslint test/javascripts`
some possible improvements to build times - use our new infrastructure based on LXC and EC2 - enable caching for bundler - use schema:load (best practice) - use --deployment (best practice when a Gemfile.lock is around) 2014-11-07 09:03:08 +08:00
Added TravisCI configuration file 2013-02-08 03:43:07 +08:00			`before_script:`
Update travis config file A few things in here: - Test on ruby 2.2 - Allow failures in all variants of RAILS_MASTER test - Use db:create task 2015-04-25 04:28:27 +08:00			`- bundle exec rake db:create db:migrate`
some possible improvements to build times - use our new infrastructure based on LXC and EC2 - enable caching for bundler - use schema:load (best practice) - use --deployment (best practice when a Gemfile.lock is around) 2014-11-07 09:03:08 +08:00
Test on rails 4.2 and master 2015-04-29 10:42:37 +08:00			`install:`
Don't try and update rails-observers Was removed from Discourse via c531f4ded5604fe0335284f12149322c16931cb7. 2016-12-24 23:39:02 +08:00			`- bash -c "if [ '$RAILS_MASTER' == '1' ]; then bundle update --retry=3 --jobs=3 arel rails seed-fu; fi"`
Test on rails 4.2 and master 2015-04-29 10:42:37 +08:00			`- bash -c "if [ '$RAILS_MASTER' == '0' ]; then bundle install --without development --deployment --retry=3 --jobs=3; fi"`
some possible improvements to build times - use our new infrastructure based on LXC and EC2 - enable caching for bundler - use schema:load (best practice) - use --deployment (best practice when a Gemfile.lock is around) 2014-11-07 09:03:08 +08:00
Fix travis script. 2016-11-30 16:41:12 +08:00			`script: "bundle exec rspec && bundle exec rake plugin:spec && bundle exec rake qunit:test['200000']"`