/*
* Jenkins Pipeline definitions for Smalltalk/X jv-branch
*
* Following stepss are defined here:
*
* * `build()` - compiles and tests all supported configurations
* No artifacts are archived
*
* * `integration()` - like `build()` steps, then all test pass on all
* configurations, archive artifacts and push staged changes to opstream
* (if approvec by the user)
*/
import com.cloudbees.plugins.credentials.CredentialsProvider;
import com.cloudbees.plugins.credentials.common.StandardUsernameCredentials;
import com.cloudbees.jenkins.plugins.sshcredentials.SSHUserPrivateKey;
/**
* Supported configurations. To overwrite, do:
*
* ...
* steps = load "steps.groovy"
* ...
* steps.configurations = [ 'BUILD_TARGET': 'mips-sgi-irix' ]
* ...
*
* Default value: see below.
*/
configurations = [
'BUILD_TARGET': ['i686-pc-mingw32' , 'x86_64-w64-mingw32', 'i686-pc-linux-gnu', 'x86_64-pc-linux-gnu' ]
]
/** Branch to build. To overwrite do:
*
* ...
* steps = load "steps.groovy"
* ...
* steps.branch = "issue-123"
* ...
*
* Default value:
* For multibranch jobs, this is the value of BRANCH_NAME environment variable.
* For normal jobs this is the value of 'Branch' field from SCM configuration.
*
* Due to Jenkins internals, the fallback to use SCM configuration must be done
* by 'user' of this variable (i.e., use `branch != null ? branch : scm.getBranch())
*/
branch = env.BRANCH_NAME
/**
* Workspace to use. To overwrite do:
*
* ...
* steps = load "steps.groovy"
* ...
* steps.workspace = "some-other-job"
* ...
*
* Default value:
* Name of current job.
*/
workspace = env.JOB_NAME
/*
* "Default" steps:
*/
def build() {
stage ( "Build" ) {
matrix ( configurations ) {
stage ( "Checkout - {${env.BUILD_TARGET}} " ) {
sshagent([ scm.getCredentialsId() ]) {
branch_to_use = branch != null ? branch : scm.getBranch()
sh """
if [ -f build.rb ]; then
hg pull --ssh ssh ${scm.getSource()}
else
hg clone --ssh ssh ${scm.getSource()} .
fi
hg up ${branch_to_use}
"""
sh "ruby build.rb --project \"stx:jv-branch\" --build-target ${env.BUILD_TARGET} update"
}
}
stage ( "Compile - {${env.BUILD_TARGET}}") {
sh "ruby build.rb --project \"stx:jv-branch\" --build-target ${env.BUILD_TARGET} compile"
}
}
}
stage ( "Test" ) {
matrix ( configurations ) {
stage ( "Test - {${env.BUILD_TARGET}}") {
/*
* Some tests requires display, so:
*
* * on *NIX hosts, launch Xvfb
* * on Windows, do nothing. Windows slave must be configured
* so it has an access to interactive window station (which
* means it has to run under user session, not as a service)
*/
if ( isUnix() ) {
wrap([$class: 'Xvfb', autoDisplayName: true, additionalOptions: '-screen 0 1024x768x24 -pixdepths 24 4 8 15 16 32', parallelBuild: true]) {
sh "ruby build.rb --project \"stx:jv-branch\" --build-target ${env.BUILD_TARGET} test"
}
} else {
sh "ruby build.rb --project \"stx:jv-branch\" --build-target ${env.BUILD_TARGET} test"
}
junit allowEmptyResults: true, testResults: "reports/*build${env.BUILD_NUMBER}*.xml"
}
}
}
}
/*
* "Integration" steps
*/
def integration() {
build()
/*
* If a single test fails, abort the steps. There's no point
* archiving a broken build.
*/
println "Smalltalk/X built, job status is: ${currentBuild.result}"
if ( currentBuild.result == 'UNSTABLE' ) {
return;
}
artifacts()
/*
* Check if there are changes to be pushed to upstream. If so,
* ask user to approve that push
*/
if ( changes() ) {
def integrate = false;
if (env.JENKINS_URL == "https://swing.fit.cvut.cz/jenkins/") {
integrate = true;
} else {
integrate = input(message: 'Integrate all staged changes to upstream?',
parameters: [
booleanParam(name: "Integrate changes",
description: 'If checked, all staged changes will be pushed to an upstream repository',
defaultValue: true)]);
}
if ( integrate ) {
push()
}
}
}
/*
* Publish built artifacts to download server
*/
def publish(dir) {
Calendar current = Calendar.getInstance()
Calendar scheduled = current.clone()
/*
* Schedule the upload to the next day, 3am in the morning
*/
scheduled.add(Calendar.DATE, 1)
scheduled.set(Calendar.HOUR_OF_DAY, 3)
build(job: 'stx_jv_publish',
parameters: [
string(name: 'job', value: env.JOB_NAME),
string(name: 'srv', value: env.STX_PUBLISH_SERVER),
string(name: 'dir', value: "${env.STX_PUBLISH_DIRECTORY}/${dir}")
],
quietPeriod: (scheduled.getTimeInMillis() - current.getTimeInMillis()) / 1000,
wait: false)
}
/*
* Utility. Return true, if there are any changes to be pushed to an upstream,
* false othervise.
*/
def changes() {
changes = false;
any ( configurations ) {
withCredentialsForUpstream() { user, pass ->
status = sh ( script: "rake \"workflow:out-upstream[${user}, ${pass}]\"", returnStatus: true)
}
changes = status == 0;
}
return changes;
}
def combine(configurations, axes = null, axis = 0, partial = new HashMap(), combinations = []) {
def axes0 = axes
if (axes0 == null) {
axes0 = configurations.keySet().toArray();
}
if ( axis < axes0.length ) {
for ( value in configurations[axes0[axis]] ) {
def combined = partial.clone()
combined[axes0[axis]] = value
combine(configurations, axes0, axis + 1, combined, combinations)
}
} else {
combinations.add(partial)
}
return combinations;
}
def matrix(configurations, block) {
def combinations = combine(configurations).toArray()
def branches = [failFast: true]
for (i = 0; i < combinations.length; i++) {
def index = i
def conf = combinations[i];
branches["${conf.BUILD_TARGET}"] = {
node ( conf.BUILD_TARGET ) {
def newEnv = []
for (k in conf.keySet()) {
newEnv.add("${k}=${conf[k]}")
}
withEnv ( newEnv ) {
ws ("workspace/${workspace}/${env.BUILD_TARGET}") {
block()
}
}
}
}
}
parallel branches
}
def any(configurations, block) {
def axes = configurations.keySet().toArray()
def conf = [:]
for (axis in axes) {
conf[axis] = configurations[axis][0]
}
node ( conf.BUILD_TARGET ) {
def newEnv = []
for (k in conf.keySet()) {
newEnv.add("${k}=${conf[k]}")
}
withEnv ( newEnv ) {
ws ("workspace/${workspace}/${env.BUILD_TARGET}") {
block()
}
}
}
}
def artifacts() {
matrix ( configurations ) {
stage ( "Artifacts - {${env.BUILD_TARGET}}") {
sh "ruby build.rb --project \"stx:jv-branch\" --build-target ${env.BUILD_TARGET} artifacts"
archiveArtifacts artifacts: "artifacts/*build${env.BUILD_NUMBER}*.zip, artifacts/*build${env.BUILD_NUMBER}*.bz2, artifacts/*build${env.BUILD_NUMBER}*.sha256", fingerprint: true//, onlyIfSuccessful: true
}
}
}
/**
* Push changes to upstream reporitory. To be called after a successful
* build. See #build()
*/
def push() {
any ( configurations ) {
stage ( "Push to upstream" ) {
withCredentialsForUpstream { user, pass ->
sh "rake \"workflow:push-upstream[${user}, ${pass}]\""
}
}
}
}
/*
* Utility. Executes given block with credentials for upstream repository.
*/
def withCredentialsForUpstream(block) {
/*
* Kludge: Upstream repositories may be on a public BitBucket
* server. To access repos on BitBucket, I (JV) don't
* want to use the same key / password as for checkouts from
* staging repositories,
*
* Therefore, also look for another credentials with ID
* `workflow:push-upstream`. If it exists, then use these to
* push to upstrem repository. If no such credentials exist,
* use standard credentials.
*
* So, here we go:
*/
def id1 = "workflow-push-upstream";
def id2 = scm.getCredentialsId();
def credentials = null;
for (StandardUsernameCredentials c : CredentialsProvider.lookupCredentials(StandardUsernameCredentials.class)) {
if (c.getId().equals(id1)) {
credentials = c;
break;
}
}
if (credentials == null) {
for (StandardUsernameCredentials c : CredentialsProvider.lookupCredentials(StandardUsernameCredentials.class)) {
if (c.getId().equals(id2)) {
credentials = c;
break;
}
}
}
println "Using upstream credentials ${credentials.getId()}: ${credentials.getDescription()}"
if (credentials instanceof SSHUserPrivateKey) {
sshagent([ credentials.getId() ]) {
// sh "rake \"workflow:push-upstream\""
block(null, null)
}
} else {
withCredentials([[$class: 'UsernamePasswordMultiBinding', credentialsId: credentials.getId(), passwordVariable: 'pass', usernameVariable: 'user']]) {
// sh "rake \"workflow:push-upstream[${user}, ${pass}]\""
block(user, pass)
}
}
}
return this;