Development cycle¶

Getting the code base¶

git clone git://git.curoverse.com/tapestry.git
cd tapestry
git submodule init
git submodule update

Running tests¶

Add these configs to your /etc/mysql/my.cnf, otherwise each "create table" will take ~0.5 seconds, and you'll wait ~3 minutes each time you run rake test before the tests start running.

• innodb_file_per_table = 0
  innodb_stats_persistent = 0
  

You can run tests in a docker container.

• Make sure /etc/mysql/my.cnf has bind-address = 0.0.0.0 or whatever is suitable.
• Create a tapestry user and a tapestry_test database on your docker host (or somewhere else reachable from a docker container).

• create database tapestry_test;
  grant all privileges on tapestry_test.* to tapestry identified by 'secretzzz';
  

• Create config/database.yml (see config/database.yml.example)
• make docker-test

Currently (March 2016) there are tests failing on master, but you should do at least this well:

• $ make docker-test
  
  [...]
  
  623 tests, 964 assertions, 228 failures, 88 errors
  Errors running test:units, test:functionals!
  Makefile:6: recipe for target 'docker-test' failed
  make: *** [docker-test] Error 1
  

Writing and contributing code¶

Start a branch like username/master or username/rails3

git branch example/rails3
git checkout example/rails3

Do something trivial like

• add app/views/pages/specimen_collection.html.erb
• add link to it in app/views/pages/home.html.erb

Commit it. If you refer to the issue# in the comment, redmine will notice this and do smart things (e.g., link to the commit from the issue page).

git add ...
git commit -m 'description of stuff (closes #123)'

Push it

git push

Generate a pull request

git request-pull e68ab19^ git@git.clinicalfuture.com:tapestry.git e68ab19

Send the pull request to someone like Ward

Merging code from other contributors/branches¶

Fetch latest code in other developers branches

git fetch

List remote branches

git branch -a

Switch to the other branch

git checkout origin/example/rails3

(git will complain about being in a detached head state, but that's fine - just don't commit to this state)

Now look around, test

git log

Switch to production branch (set up with git checkout --track origin/rails3 if you haven't already) and make sure it's up-to-date

git checkout rails3
git pull origin rails3

See what hasn't been merged

git cherry -v rails3 origin/example/rails3

Cherry-pick a commit

git cherry-pick e68ab19

Or merge the entire remote tree (this will fetch and merge but not yet commit)

git merge --no-commit origin/tomc/rails3

When you're happy,

git commit

Or rewind

git reset --hard rails3

Finally, push

git push

Deploying to my-dev¶

Basics

sudo gem install capistrano

Make sure "ssh root@www-dev.sum" puts you in root@www-dev and that you have an SSH agent available. You might need in ~/.ssh/config:

Host www-dev.sum
  HostName www-dev

Check out production branch

git checkout rails3

Deploy to www-dev

cap deploy

The capistrano task will run any required db migrations automatically. If you ever need to run them manually, you can do this (on www-dev):

ssh root@www-dev.sum sh -c "'cd /var/www/my-dev.personalgenomes.org/current && RAILS_ENV=staging rake db:migrate'"

Deploying to production server¶

Make sure "ssh root@www-prod.sum" puts you in root@www-prod and that you have a working SSH agent.

Check out production branch

git checkout rails3

Deploy to www-prod

cap deploy -f Capfile.production

The capistrano task will run any required db migrations automatically. If you ever need to run them manually, you can do this (on www-prod):

ssh root@www-prod.sum sh -c "'cd /var/www/my.personalgenomes.org/current && RAILS_ENV=production rake db:migrate'"