The Travis Client

The travis gem includes both a command line client and a Ruby library to interface with a Travis CI service. Both work with, or any custom Travis CI setup you might have.

Table of Contents

Command Line Client

There are three types of commands: Non-API Commands, General API Commands and Repository Commands. All commands take the form of travis COMMAND [ARGUMENTS] [OPTIONS]. You can get a list of commands by running help.

Non-API Commands

Every Travis command takes three global options:

-h, --help                       Display help
-i, --[no-]interactive           be interactive and colorful
-E, --[no-]explode               don't rescue exceptions

The --help option is equivalent to running travis help COMMAND.

The --interactive options determines wether to include additional information and colors in the output or not (except on Windows, we never display colors on Windows, sorry). If you don't set this option explicitly, you will run in interactive mode if you invoke the command directly in a shell and in non-interactive mode if you pipe it somewhere.

You probably want to use --explode if you are working on a patch for the Travis client, as it will give you the Ruby exception instead of a nice error message.


The help command will inform you about the arguments and options that the commands take, for instance:

$ travis help help
Usage: travis help [command] [options]
    -h, --help                       Display help
    -i, --[no-]interactive           be interactive and colorful
    -E, --[no-]explode               don't rescue exceptions

Running help without a command name will give you a list of all available commands.


As you might have guessed, this command prints out the client's version.

General API Commands

API commands inherit all options from Non-API Commands.

Additionally, every API command understands the following options:

-e, --api-endpoint URL           Travis API server to talk to
    --pro                        short-cut for --api-endpoint ''
    --org                        short-cut for --api-endpoint ''
-t, --token [ACCESS_TOKEN]       access token to use
    --debug                      show API requests
    --adapter ADAPTER            Faraday adapter to use for HTTP requests

By default, General API Commands will talk to You can change this by supplying --pro for or --api-endpoint with your own endpoint. Note that all Repository Commands will try to figure out the API endpoint to talk to automatically depending on the project's visibility on GitHub.

You can supply an access token via --token if you want to make an authenticated call. If you don't have an access token stored for the API endpoint, it will remember it for subsequent requests. Keep in mind, this is not the "Travis token" used when setting up GitHub hooks (due to security). You probably don't have an access token handy right now. Don't worry, usually you won't use this option but instead just do a travis login.

The --debug option will print HTTP requests to STDERR. Like --explode, this is really helpful when contributing to this project.

There are many libraries out there to do HTTP requests in Ruby. You can switch amongst common ones with --adapter:

$ travis show --adapter net-http
$ gem install excon
$ travis show --adapter excon


Running travis console gives you an interactive Ruby session with all the entities imported into global namespace.

But why use this over just irb -r travis? For one, it will take care of authentication, setting the correct endpoint, etc, and it also allows you to pass in --debug if you are curious as to what's actually going on.

$ travis console
>> User.current
=> #<User: rkh>
>> Repository.find('sinatra/sinatra')
=> #<Repository: sinatra/sinatra>
>> _.last_build
=> #<Travis::Client::Build: sinatra/sinatra#360>


Just prints out the API endpoint you're talking to.

$ travis endpoint
API endpoint:

Handy for using it when working with shell scripts:

$ curl "$(travis endpoint)/docs" > docs.html


The login command will, well, log you in. That way, all subsequent commands that run against the same endpoint will be authenticated.

$ travis login
We need your GitHub login to identify you.
This information will not be sent to Travis CI, only to GitHub.
The password will not be displayed.

Try running with --github-token or --auto if you don't want to enter your password anyways.

Username: rkh
Password: *******************

Successfully logged in!

As you can see above, it will ask you for your GitHub user name and password, but not send these to Travis CI. Instead, it will use them to create a GitHub API token, show the token to Travis, which then on its own checks if you really are who you say you are, and gives you an access token for the Travis API in return. The client will then delete the GitHub token again, just to be sure. But don't worry, all that happens under the hood and fully automatic.

If you don't want it to send your credentials to GitHub, you can create a GitHub token on your own and supply it via --github-token. In that case, the client will not delete the GitHub token (as it can't, it needs your password to do this). Travis CI will not store the token, though - after all, it already should have a valid token for you in the database.

A third option is for the really lazy: --auto. In this mode the client will try to find a GitHub token for you and just use that. This will only work if you have a global GitHub token stored in your .netrc. If you haven't heard of this, it's worth looking into in general. Again: Travis CI will not store that token.


This is really helpful both when working on this client and when exploring the Travis API. It will simply fire a request against the API endpoint, parse the output and pretty print it. Keep in mind that the client takes care of authentication for you:

$ travis raw /repos/travis-ci/travis
   "description"=>"Travis CI Client (CLI and Ruby library)",

Use --json if you'd rather prefer the output to be JSON.


Usage: travis sync [options]
    -h, --help                       Display help
    -i, --[no-]interactive           be interactive and colorful
    -E, --[no-]explode               don't rescue exceptions
    -e, --api-endpoint URL           Travis API server to talk to
        --pro                        short-cut for --api-endpoint ''
        --org                        short-cut for --api-endpoint ''
    -t, --token [ACCESS_TOKEN]       access token to use
        --debug                      show API requests
    -c, --check                      only check the sync status
    -b, --background                 will trigger sync but not block until sync is done
    -f, --force                      will force sync, even if one is already running

Sometimes the infos Travis CI has about users and repositories become out of date. If that should happen, you can manually trigger a sync:

$ travis sync
synchronizing: ........... done

The command blocks until the synchronization is done. You can avoid that with --background:

$ travis sync --background
starting synchronization

If you just want to know if your account is being synchronized right now, use --check:

$ travis sync --check
rkh is currently syncing


In order to use the Ruby library you will need to obtain an access token first. To do this simply run the travis login command. Once logged in you can check your token with travis token:

$ travis token
Your access token is super-secret

You can use that token for instance with curl:

$ curl -H "Authorization: token $(travis token)"
{"login":"rkh","name":"Konstantin Haase","email":"[email protected]","gravatar_id":"5c2b452f6eea4a6d84c105ebd971d2a4","locale":"en","is_syncing":false,"synced_at":"2013-01-21T20:31:06Z"}

Note that if you just need it for looking at API payloads, that we also have the raw command.


It's just a tiny feature, but it allows you to take a look at repositories that have recently seen some action (ie the left hand sidebar on

$ travis whatsup
mysociety/fixmystreet started: #154
eloquent/typhoon started: #228
Pajk/apipie-rails started: #84
qcubed/framework failed: #21


This command is useful to verify that you're in fact logged in:

$ travis whoami
You are rkh (Konstantin Haase)

Again, like most other commands, goes well with shell scripting:

$ git clone "$(travis whoami)/some_project"

Repository Commands

-h, --help                       Display help
-i, --[no-]interactive           be interactive and colorful
-E, --[no-]explode               don't rescue exceptions
-e, --api-endpoint URL           Travis API server to talk to
    --pro                        short-cut for --api-endpoint ''
    --org                        short-cut for --api-endpoint ''
-t, --token [ACCESS_TOKEN]       access token to use
    --debug                      show API requests
-r, --repo SLUG

Repository commands have all the options General API Commands have.

Additionally, you can specify the Repository to talk to by providing --repo owner/name. However, if you invoke the command inside a clone of the project, the client will figure out this option on its own. Note that it uses the git remote "origin" to do so.

It will also automatically pick Travis Pro if it is a private project. You can of course override this decission with --pro, --org or --api-endpoint URL


If you want to turn of a repository temporarily or indefinitely, you can do so with the disable command:

$ travis disable
travis-ci/travis: disabled :(


With the enable command, you can easily activate a project on Travis CI:

$ travis enable
travis-ci/travis: enabled :)

It even works when enabling a repo Travis didn't know existed by triggering a sync:

$ travis enable -r rkh/test
repository not known to Travis CI (or no access?)
triggering sync: ............. done
rkh/test: enabled

If you don't want the sync to be triggered, use --skip-sync.


Usage: travis encrypt [args..] [options]
    -h, --help                       Display help
    -i, --[no-]interactive           be interactive and colorful
    -E, --[no-]explode               don't rescue exceptions
    -e, --api-endpoint URL           Travis API server to talk to
        --pro                        short-cut for --api-endpoint ''
        --org                        short-cut for --api-endpoint ''
    -t, --token [ACCESS_TOKEN]       access token to use
        --debug                      show API requests
    -r, --repo SLUG
        --add [KEY]                  adds it to .travis.yml under KEY (default:
    -s, --[no-]split                 treat each line as a separate input

This command is useful to encrypt environment variables or deploy keys for private dependencies.

$ travis encrypt FOO=bar
Please add the following to your .travis.yml file:

  secure: "gSly+Kvzd5uSul15CVaEV91ALwsGSU7yJLHSK0vk+oqjmLm0jp05iiKfs08j\n/Wo0DG8l4O9WT0mCEnMoMBwX4GiK4mUmGdKt0R2/2IAea+M44kBoKsiRM7R3\n+62xEl0q9Wzt8Aw3GCDY4XnoCyirO49DpCH6a9JEAfILY/n6qF8="

Pro Tip™: You can add it automatically by running with --add.

For deploy keys, it is really handy to pipe them into the command:

$ cat id_rsa | travis encrypt

Another use case for piping files into it: If you have a file with sensitive environment variables, like foreman's .env file, you can add tell the client to encrypt every line separately via --split:

$ cat .env | travis encrypt --split
Please add the following to your .travis.yml file:

  secure: "KmMdcwTWGubXVRu93/lY1NtyHxrjHK4TzCfemgwjsYzPcZuPmEA+pz+umQBN\n1ZhzUHZwDNsDd2VnBgYq27ZdcS2cRvtyI/IFuM/xJoRi0jpdTn/KsXR47zeE\nr2bFxRqrdY0fERVHSMkBiBrN/KV5T70js4Y6FydsWaQgXCg+WEU="
  secure: "jAglFtDjncy4E3upL/RF0ZOcmJ2UMrqHFCLQwU8PBdurhTMBeTw+IO6cXx5z\nU5zqvPYo/ghZ8mMuUhvHiGDM6m6OlMP7+l10VTxH1CoVew2NcQvRdfK3P+4S\nZJ43Hyh/ZLCjft+JK0tBwoa3VbH2+ZTzkRZQjdg54bE16C7Mf1A="

Pro Tip™: You can add it automatically by running with --add.

As suggested, the client can also add them to your .travis.yml for you:

$ travis encrypt FOO=bar --add

This will by default add it as global variables for every job. You can also add it as matrix entries by providing a key:

$ travis encrypt FOO=bar --add env.matrix


Usage: travis history [options]
    -h, --help                       Display help
    -i, --[no-]interactive           be interactive and colorful
    -E, --[no-]explode               don't rescue exceptions
    -e, --api-endpoint URL           Travis API server to talk to
        --pro                        short-cut for --api-endpoint ''
        --org                        short-cut for --api-endpoint ''
    -t, --token [ACCESS_TOKEN]       access token to use
        --debug                      show API requests
    -r, --repo SLUG
    -a, --after BUILD                Only show history after a given build number
    -p, --pull-request NUMBER        Only show history for the given Pull Request
    -b, --branch BRANCH              Only show history for the given branch
    -l, --limit LIMIT                Maximum number of history items
        --[no-]all                   Display all history items

You can check out what the recent builds look like:

$ travis history
#77 passed:   master fix name clash
#76 failed:   master Merge pull request #11 from travis-ci/rkh-show-logs-history
#75 passed:   rkh-debug what?
#74 passed:   rkh-debug all tests pass locally and on the travis vm I spin up :(
#73 failed:   Pull Request #11 regenerate gemspec
#72 passed:   rkh-show-logs-history regenerate gemspec
#71 failed:   Pull Request #11 spec fix for (older) rubinius
#70 passed:   rkh-show-logs-history spec fix for (older) rubinius
#69 failed:   Pull Request #11 strange fix for rubinius
#68 failed:   rkh-show-logs-history strange fix for rubinius

By default, it will display the last 10 builds. You can limit (or extend) the number of builds with --limit:

$ travis history --limit 2
#77 passed:   master fix name clash
#76 failed:   master Merge pull request #11 from travis-ci/rkh-show-logs-history

You can use --after to display builds after a certain build number (or, well, before, but it's called after to use the same phrases as the API):

$ travis history --limit 2 --after 76
#75 passed:   rkh-debug what?
#74 passed:   rkh-debug all tests pass locally and on the travis vm I spin up :(

You can also limit the history to builds for a certain branch:

$ travis history --limit 3 --branch master
#77 passed:   master fix name clash
#76 failed:   master Merge pull request #11 from travis-ci/rkh-show-logs-history
#57 passed:   master Merge pull request #5 from travis-ci/hh-multiline-encrypt

Or a certain Pull Request:

$ travis history --limit 3 --pull-request 5
#56 passed:   Pull Request #5 Merge branch 'master' into hh-multiline-encrypt
#49 passed:   Pull Request #5 improve output
#48 passed:   Pull Request #5 let it generate accessor for line splitting automatically


Given a job number, logs simply prints out that job's logs.

$ travis logs 77.1
[... more logs ...]
Your bundle is complete! Use `bundle show [gemname]` to see where a bundled gem is installed.
$ bundle exec rake
/home/travis/.rvm/rubies/ruby-1.8.7-p371/bin/ruby -S rspec spec -c
Faraday: you may want to install system_timer for reliable timeouts

Finished in 6.48 seconds
163 examples, 0 failures

Done. Build script exited with: 0


Opens the project view in the Travis CI web interface. If you pass it a build or job number, it will open that specific view:

$ travis open

If you just want the URL printed out instead of opened in a browser, pass --print.

If instead you want to open the repository, compare or pull request view on GitHub, use --github.

$ travis open 56 --print --github
web view:


This command will restart the latest build:

$ travis restart
build #85 has been restarted

You can also restart any build by giving a build number:

$ travis restart 57
build #57 has been restarted

Or a single job:

$ travis restart 57.1
job #57.1 has been restarted


Displays general infos about the latest build:

$ travis show
Build #77: fix name clash
State:         passed
Type:          push
Compare URL:
Duration:      5 min 51 sec
Started:       2013-01-19 19:00:49
Finished:      2013-01-19 19:02:17

#77.1 passed:    45 sec         rvm: 1.8.7
#77.2 passed:    50 sec         rvm: 1.9.2
#77.3 passed:    45 sec         rvm: 1.9.3
#77.4 passed:    46 sec         rvm: 2.0.0
#77.5 failed:    1 min 18 sec   rvm: jruby (failure allowed)
#77.6 passed:    1 min 27 sec   rvm: rbx

Any other build:

$ travis show 1
Build #1: add .travis.yml
State:         failed
Type:          push
Compare URL:
Duration:      3 min 16 sec
Started:       2013-01-13 23:15:22
Finished:      2013-01-13 23:21:38

#1.1 failed:     21 sec         rvm: 1.8.7
#1.2 failed:     34 sec         rvm: 1.9.2
#1.3 failed:     24 sec         rvm: 1.9.3
#1.4 failed:     52 sec         rvm: 2.0.0
#1.5 failed:     38 sec         rvm: jruby
#1.6 failed:     27 sec         rvm: rbx

Or a job:

$ travis show 77.3
Job #77.3: fix name clash
State:         passed
Type:          push
Compare URL:
Duration:      45 sec
Started:       2013-01-19 19:00:49
Finished:      2013-01-19 19:01:34
Allow Failure: false
Config:        rvm: 1.9.3


Usage: travis status [options]
    -h, --help                       Display help
    -i, --[no-]interactive           be interactive and colorful
    -E, --[no-]explode               don't rescue exceptions
    -e, --api-endpoint URL           Travis API server to talk to
        --pro                        short-cut for --api-endpoint ''
        --org                        short-cut for --api-endpoint ''
    -t, --token [ACCESS_TOKEN]       access token to use
        --debug                      show API requests
    -r, --repo SLUG
    -x, --[no-]exit-code             sets the exit code to 1 if the build failed
    -q, --[no-]quiet                 does not print anything
    -p, --[no-]fail-pending          sets the status code to 1 if the build is pending

Outputs a one line status message about the project's last build. With -q that line will even not be printed out. How's that useful? Combine it with -x and the exit code will be 1 if the build failed, with -p and it will be 1 for a pending build.

$ travis status -qpx && cap deploy

Ruby Library

There are two approaches of using the Ruby library, one straight forward with one global session:

require 'travis'

rails = Travis::Repository.find('rails/rails')
puts "oh no" unless

And one where you have to instantiate your own session:

require 'travis/client'

client =
rails  = client.repo('rails/rails')
puts "oh no" unless

For most parts, those are pretty much the same, the entities you get back look the same, etc, except one offers nice constants as part of the API, the other doesn't. In fact the "global" session style uses Travis::Client internally.

So, which one to choose? The global style has one session, whereas with the client style, you have one session per client instance. Each session has it's own cache and identity map. This might matter for log running processes. If you use a new session for separate units of work, you can be pretty sure to not leak any objects. On the other hand using the constants or reusing the same session might save you from unnecessary HTTP requests.

In either way, if you should use the first approach or long living clients, here is how you make sure not to have stale data around:


Note that this will still keep the identity map around, it will only drop all attributes. To clear the identity map, you can use the clear_cache! method. However, if you do that, you should not keep old instances of any entities (like repositories, etc) around.


Authentication is pretty easy, you just need to set an access token:

require 'travis'

Travis.access_token = "..."
puts "Hello #{}!"

Or with your own client instance:

require 'travis/client'

client =
puts "Hello #{}"

See the token command for obtaining the access token used by the CLI.

If you don't have an access token for Travis CI, you can use a GitHub access token to get one:

require 'travis'

puts "Hello #{}!"

Travis CI will not store that token.

Using Pro

Using the library with private projects pretty much works the same, except you use Travis::Pro.

Keep in mind that you need to authenticate.

require 'travis/pro'

Travis::Pro.access_token = '...'
user = Travis::Pro::User.current

puts "Hello #{}!"


Entities are like the models in the Travis Client land. They keep the data and it's usually them you talk to if you want something. They are pretty much normal Ruby objects.

The Travis session will cache all entities, so don't worry about loading the same one twice. Once you got a hold of one, you can easily reload it at any time if you want to make sure the data is fresh:

rails = Travis::Repository.find('rails/rails')
sleep 1.hour

The travis gem supports lazy and partial loading, so if you want to make sure you have all the data, just call load.


This is not something you should usually do, as partial loading is actually your friend (keeps requests to a minimum).

Stateful Entities

Repositories, Builds and Jobs all are basically state machines, which means the implement the following methods:

require 'travis'
build = Travis::Repository.find('rails/rails').last_build

p build.canceled?
p build.created?
p build.errored?
p build.failed?
p build.finished?
p build.passed?
p build.pending?
p build.queued?
p build.running?
p build.started?
p build.successful?
p build.unsuccessful?
p build.yellow?
p build.color

Builds and jobs also have a state method. For repositories, use last_build.state.


Repositories are probably one of the first entities you'll load. It's pretty straight forward, too.

require 'travis'

Travis::Repository.find('rails/rails')            # find by slug
Travis::Repository.find(891)                      # find by id
Travis::Repository.find_all(owner_name: 'rails')  # all repos in the rails organization
Travis::Repository.current                        # repos that see some action right now

Once you have a repository, you can for instance encrypt some strings with its private key:

require 'travis'

puts repo.encrypt('FOO=bar')

Repositories are stateful.

You can enable or disable a repository with the methods that go by the same name.

system "push all the things"

If you want to enable a new project, you might have to do a sync first.


You could load a build by its id using Travis::Build.find. But most of the time you won't have the id handy, so you'd usually start with a repository.

require 'travis'
rails = Travis::Repository.find('rails/rails')

rails.last_build               # the latest build
rails.recent_builds            # the last 20 or so builds (don't rely on that number)
rails.builds(after_number: 42) # the last 20 or so builds *before* 42                # build with the number 42 (not the id!)
rails.builds                   # Enumerator for #each_build

# this will loop through all builds
rails.each_build do |build|
  puts "#{build.number}: #{build.state}"

# this will loop through all builds before build 42
rails.each_build(after_number: 42) do |build|
  puts "#{build.number}: #{build.state}"

Note that each_build (and thus builds without and argument) is lazy and uses pagination, so you can safely do things like this:

build = rails.builds.detect { |b| b.failed? }
puts "Last failing Rails build: #{build.number}"

Without having to load more than 6000 builds.

You can restart a build, if the current user has sufficient permissions on the repository:



Jobs behave a lot like builds, and similar to them, you probably don't have the id ready. You can get the jobs from a build: do |job|
  puts "#{job.number} took #{job.duration} seconds"

If you have the job number, you can also reach a job directly from the repository:


Like builds, you can also restart singe jobs:



The artifacts you usually care for are probably logs. You can reach them directly from a build:

require 'travis'

repo = Travis::Repository.find('travis-ci/travis')
job  =
puts job.log.body

If you plan to print our the body, be aware that it might contain malicious escape codes. For this reason, we added colorized_body, which removes all the unprintable characters, except for ANSI color codes, and clean_body which also removes the color codes.

puts job.log.colorized_body


The only user you usually get access to is the currently authenticated one.

require 'travis'

Travis.access_token = '...'
user = Travis::User.current

puts "Hello, #{user.}! Or should I call you... #{}!?"

If some data gets out of sync between GitHub and Travis, you can use the user object to trigger a new sync.



Commits cannot be loaded directly. They come as a byproduct of jobs and builds.

require 'travis'

repo   = Travis::Repository.find('travis-ci/travis')
commit = repo.last_build.commit

puts "Last tested commit: #{commit.short_sha} on #{commit.branch} by #{commit.author_name} - #{commit.subject}"


If a worker is running something, it will reference a job and a repository. Otherwise the values will be nil.

require 'travis'
workers = Travis::Worker.find_all

workers.each do |worker|
  puts "#{}: #{} - #{worker.state} - #{worker.repository.slug if worker.repository}"

Dealing with Sessions

Under the hood the session is where the fun is happening. Most methods on the constants and entities just wrap methods on your session, so you don't have to pass the session around all the time or even see it if you don't want to.

There are two levels of session methods, the higher level methods from the Travis::Client::Methods mixin, which are also available from Travis, Travis::Pro or any custom Namespace.

require 'travis/client/session'
session =

session.access_token = "secret_token"           # access token to use
session.api_endpoint = "http://localhost:3000/" # api endpoint to talk to
session.github_auth("github_token")             # log in with a github token
session.repos(owner_name: 'travis-ci')          # all travis-ci/* projects
session.repo('travis-ci/travis')                # this project
session.repo(409371)                            # same as the one above                          # build with id 4266036
session.job(4266037)                            # job with id 4266037
session.artifact(42)                            # artifact with id 42
session.log(42)                                 # same as above
session.user                                    # the current user, if logged in
session.restart(         # restart some build

You can add these methods to any object responding to session via said mixin.

Below this, there is a second API, close to the HTTP level:

require 'travis/client/session'
session =

session.instrument do |description, block|
  time =
  puts "#{description} took #{ - time} seconds"

session.connection =

session.get_raw('/repos/rails/rails') # => {"repo" => {"id" => 891, "slug" => "rails/rails", ...}}
session.get('/repos/rails/rails')     # => {"repo" => #<Travis::Client::Repository: rails/rails>}
session.headers['Foo'] = 'Bar'        # send a custom HTTP header with every request

rails = session.find_one(Travis::Client::Repository, 'rails/rails')

session.find_many(Travis::Client::Repository)  # repositories with the latest builds
session.find_one_or_many(Travis::Client::User) # the current user (you could also use find_one here)

session.reset(rails)  # lazy reload

session.clear_cache   # empty cached attributes
session.clear_cache!  # empty identity map

Using Namespaces

Travis and Travis::Pro are just two different namespaces for two different Travis sessions. A namespace is a Module, exposing the higher level session methods. It also has a dummy constant for every entity, wrapping find_one (aliased to find) and find_many (aliased to find_all) for you, so you don't have to keep track of the session or hand in the entity class. You can easily create your own namespace:

require 'travis/client'
MyTravis ="http://localhost:3000")

MyTravis.access_token = "..."

Since namespaces are Modules, you can also include them.

require 'travis/client'

class MyTravis



Make sure you have at least Ruby 1.8.7 (1.9.3 recommended) installed. Then run:

$ gem install travis --no-rdoc --no-ri

Upgrading from travis-cli

If you have the old travis-cli gem installed, you should gem uninstall travis-cli, just to be sure, as it ships with an executable that is also named travis.

Version History

Not yet released

  • add --adapter to API endpoints
  • added branch to show
  • fix bug where colors were not used if stdin is a pipe
  • make encrypt options --split and --add work together properly
  • better handling of missing or empty .travis.yml when running encrypt --add
  • fix broken example code
  • no longer require network connection to automatically detect repository slug
  • add worker support to the ruby library
  • adjust artifacts/logs code to upstream api changes

v1.1.3 (January 26, 2013)

  • use persistent HTTP connections (performance for commands with multiple api requests)
  • include round trip time in debug output

v1.1.2 (January 24, 2013)

  • token command
  • no longer wrap $stdin in delegator (caused bug on some Linux systems)
  • correctly detect when running on Windows, even on JRuby

v1.1.1 (January 22, 2013)

  • Make pry a runtime dependency rather than a development dependency.

v1.1.0 (January 21, 2013)

  • New commands: console, status, show, logs, history, restart, sync, enable, disable, open and whatsup.
  • --debug option for all API commands.
  • --split option for encrypt.
  • Fix --add option for encrypt (was naming key secret instead of secure).
  • First class representation for builds, commits and jobs in the Ruby library.
  • Print warning when running "encrypt owner/project data", as it's not supported by the new client.
  • Improved documentation.

v1.0.3 (January 15, 2013)

  • Fix -r slug for repository commands. (#3)

v1.0.2 (January 14, 2013)

  • Only bundle CA certs needed to verify Travis CI and GitHub domains.
  • Make tests pass on Windows.

v1.0.1 (January 14, 2013)

  • Improve encrypt --add behavior.

v1.0.0 (January 14, 2013)

  • Fist public release.
  • Improved documentation.

v1.0.0pre2 (January 14, 2013)

  • Added Windows support.
  • Suggestion to run travis login will add --org if needed.

v1.0.0pre (January 13, 2013)

  • Initial public prerelease.