You can find both reference documentation and developer documentation in the following location:
The following guide is mainly targeted towards a Linux or Mac OSX development machine. If you are on Windows, we recommend to take a look at Getting started with fabric8-cluster development on Windows.
You need to install:
-
go
(>= v1.8) -
git
-
mercurial
-
make
Run the following command to find out your Go version.
$ go version
You must at least have Go version 1.8.
See Fetch dependencies to see an explanaition on how we deal with dependencies.
This project uses dep as a package manager for Go.
Running the make deps
command will install dep
in $GOPATH/bin
if it’s not already available on your system.
Assuming you have Go installed and configured (have $GOPATH
setup) here is
how to build.
Check out the code
$ git clone https://github.com/fabric8-services/fabric8-cluster $GOPATH/src/github.com/fabric8-services/fabric8-cluster
Like most other projects, this one depends on various other projects that need to be downloaded.
We also generate some code from design files that shall make it into our final artifacts.
To fetch the dependencies, generate code and finally build the project you can
type make
in a freshly clone repository of this project.
$ cd $GOPATH/src/github.com/fabric8-services/fabric8-cluster $ make
If you are getting the following error
[ERROR] Failed to set references: Unable to update checked out version (Skip to cleanup) make: *** [vendor] Error 1
Try out this
$ glide cache-clear $ make clean deps $ make build
There is no need to fetch the dependencies, or re-generate code every time you
want to compile. That’s why we offer special make
targets for these topics:
This will download all the dependencies for this project inside a directory
called vendor
. This way we can ensure that every developer and our CI system
is using the same version.
$ cd $GOPATH/src/github.com/fabric8-services/fabric8-cluster $ make deps
For dependency management of go
packages we use dep
.
The file Gopkg.toml
contains all dependencies. If you want to understand the format for this file, look here.
You need to run this command if you just checked out the code and later if you’ve modified the designs.
$ cd $GOPATH/src/github.com/fabric8-services/fabric8-cluster $ make generate
If you want to just build the Auth server and client, run make build
.
$ cd $GOPATH/src/github.com/fabric8-services/fabric8-cluster $ make build
This removes all downloaded dependencies, all generated code and compiled artifacts.
$ cd $GOPATH/src/github.com/fabric8-services/fabric8-cluster $ make clean
Here’s how to run all available tests. All tests will check all Go packages
except those in the vendor/
directory.
Make sure you have docker and docker-compose available.
Setting up test environment - make integration-test-env-prepare
Tear test environment down - make integration-test-env-tear-down
unit-tests |
Unit tests have the minimum requirement on time and environment setup. $ cd $GOPATH/src/github.com/fabric8-services/fabric8-cluster $ make test-unit |
integration-tests |
Integration tests demand more setup (i.e. the PostgreSQL DB must be already
running) and probably time. We recommend that you use $ cd $GOPATH/src/github.com/fabric8-services/fabric8-cluster $ make test-integration |
all |
To run both, the unit and the integration tests you can run $ cd $GOPATH/src/github.com/fabric8-services/fabric8-cluster $ make test-all |
By default, test data is removed from the database after each test, unless the AUTH_CLEAN_TEST_DATA
environment variable is set to false
. This can be particularily useful to run queries on the test data after a test failure, in order to understand why the result did not match the expectations.
Also, all SQL queries can be displayed in the output if the AUTH_ENABLE_DB_LOGS
environment variable is set to `true. Beware that this can be very verbose, though ;)
To visualize the coverage of unit, integration, or all tests you can run these commands:
-
$ make coverage-unit
-
$ make coverage-integration
-
$ make coverage-all
Note
|
If the tests (see Tests) have not yet run, or if the sources have changed since the last time the tests ran, they will be re-run to produce up to date coverage profiles. |
Each of the above tests (see Tests) produces a coverage profile by default. Those coverage files are available under
tmp/coverage/<package>/coverage.<test>.mode-<mode>
Here’s how the <placeholders> expand
<package>
|
something like |
<test>
|
|
<mode>
|
Sets the mode for coverage analysis for the packages being tested.
Possible values for
|
In addition to all individual coverage information for each package, we also create three more files:
tmp/coverage.unit.mode-<mode>
|
This file collects all the coverage profiles for all unit tests. |
tmp/coverage.integration.mode-<mode>
|
This file collects all the coverage profiles for all integration tests. |
tmp/coverage.mode-<mode>
|
This file is the merge result of the two afore mentioned files and thus gives coverage information for all tests. |
Start up dependent docker services using docker-compose
and runs auto reload on source change tool fresh
.
$ cd $GOPATH/src/github.com/fabric8-services/fabric8-cluster $ make dev
The above steps start the API Server on port 8087.
Test out the build by executing CLI commands in a different terminal.
Note
|
The CLI needs the API Server which was started on executing make dev to be up and running. Please do not kill the process. Alternatively if you haven’t run make dev you could just start the server by running ./bin/cluster .
|
Generate a token for future use.
./bin/cluster-cli generate token -H localhost:8087 --pp
You should get Token in response, save this token in your favourite editor as you need to use this token for POST API calls
The database are kept in a docker container that gets reused between restarts. Thus restarts will not clear out the database.
To clear out the database kill the database like this:
$ docker kill fabric8auth_db_1 && docker rm fabric8auth_db_1
In case you have mulitple fabric8*
running use docker ps
to locate the container name.
See the following README in the minishift directory for instructions on running fabric8-cluster in minishift: