riboseinc / aws-sqs-plugin Goto Github PK

As title. Also detail syntax to subscribe to one and multiple branches.

Do integration test except connecting to real SQS, by using sqsmock

Fixtures

• Trigger on Branch without Wildcard
• Trigger on Branch with Prefix Wildcard
• Trigger on Branch with Suffix Wildcard
• Trigger on all Branches ("*" or "**")

Update all using libs to the latest versions

Jenkins' AWS SQS plugin today triggers all Jenkins jobs that subscribe to the SQS queue, but does not perform filtering on Git branches.

The problem occurs in the following scenario:

• a CodeCommit git repo has two branches, "master" and "development"
• a SQS queue is created for this repo to monitor changes to it
• there are 2 Jenkins jobs, 1 for building "master" and 1 for building "development"
• these two Jenkins jobs are configured to build using the Jenkins SQS plugin subscribing to the SQS queue.

Here, any code push to the "development" branch, will trigger both "master" and "development" Jenkins jobs, which is very bad especially when there are multiple branches being tracked.

The reason for this is, the Jenkins AWS SQS plugin is not smart enough to trigger the "right" jobs -- the SQS message in fact already tells you what branch has been changed. If we can make the plugin smarter, and provide an extra field when enabling the plugin for the job called "Subscribed branches" (which can default to the branch referred to in the build steps), the plugin only needs to trigger the jobs that match the subscribed branch.

Wildcard should be supported like it is in Jenkin Git Plugin

• * is recognized as a wildcard
• ** is recognized as wildcard that includes the separator '/'

Therefore, origin/branches* would match origin/branches-foobut not origin/branches/foo, while origin/branches** would match both origin/branches-foo and origin/branches/foo

https://github.com/jenkinsci/plugin-pom

Currently the plugin requires 1 SQS queue for 1 CodeCommit repo. It would be much more efficient if we only use 1 SQS queue for N CodeCommit repos.

This will require adding to the Job page, a textbox after the SQS queue selector but before the branch textbox, used to match the CodeCommit repo. The default CodeCommit repo selected should be the one used in the Git plugin for the job. Perhaps there is a way to determine which CodeCommit repos publish to a selected SQS queue.

Getting 404 when query the log

404 Not Found

Stapler processed this HTTP request as follows, but couldn't find the resource to consume the request

-> evaluate(<hudson.model.Hudson@200cb5de> :hudson.model.Hudson,"/job/testjenkins-sqstest/SQSActivityLog")
-> evaluate(((StaplerProxy)<hudson.model.Hudson@200cb5de>).getTarget(),"/job/testjenkins-sqstest/SQSActivityLog")
-> evaluate(<hudson.model.Hudson@200cb5de>.getJob("testjenkins-sqstest"),"/SQSActivityLog")
-> evaluate(<hudson.model.FreeStyleProject@56e5e666[testjenkins-sqstest]> :hudson.model.FreeStyleProject,"/SQSActivityLog")
-> evaluate(<hudson.model.FreeStyleProject@56e5e666[testjenkins-sqstest]>.getDynamic("SQSActivityLog",...),"")
-> evaluate(<io.relution.jenkins.awssqs.SQSTrigger$SQSTriggerPollingAction@19b2cd> :io.relution.jenkins.awssqs.SQSTrigger$SQSTriggerPollingAction,"")
-> No matching rule was found on <io.relution.jenkins.awssqs.SQSTrigger$SQSTriggerPollingAction@19b2cd> for ""
<io.relution.jenkins.awssqs.SQSTrigger$SQSTriggerPollingAction@19b2cd> has the following URL mappings, in the order of preference:

TOKEN.groovy for url=/TOKEN
VIEW.jelly for url=/VIEW
io.relution.jenkins.awssqs.SQSTrigger$SQSTriggerPollingAction.getDisplayName() for url=/displayName/...
io.relution.jenkins.awssqs.SQSTrigger$SQSTriggerPollingAction.getLog() for url=/log/...
io.relution.jenkins.awssqs.SQSTrigger$SQSTriggerPollingAction.getOwner() for url=/owner/...
io.relution.jenkins.awssqs.SQSTrigger$SQSTriggerPollingAction.getUrlName() for url=/urlName/...
io.relution.jenkins.awssqs.SQSTrigger$SQSTriggerPollingAction.getIconFileName() for url=/iconFileName/...
java.lang.Object.getClass() for url=/class/...