aegershman / pivotal-tracker-resource Goto Github PK

https://github.com/pivotal-cf/pivnet-resource

My primary workflow is creating a story based on the name of a pivnet resource. In order to create that name, you have to use a task to pull the product's version from it's metadata. It would be nice if there were better integration with the pivnet product, where I could simply feed the pivnet-product in as a param, and could get access to it's properties; e.g., product_name, product_slug, product_version, etc. etc.

owner_ids: [3016456] isn't the most appealing way to assign stories to users. I haven't explored the pivotal tracker API enough to know if it's even possible to reference owner_ids by username/email, or if the pivotal tracker client being used can do this already. Regardless, I don't want to reference owners through their database ID.

This project should be versioned. Pretty self explanatory. Gotta get that done.

Currently not unit tests at all. That's ... not great.

Should the ref be the storyID of what was created/updated? Should the ref be the last_changed time of the story? Figure this out.

Would be nice to have tighter control over when a story is created/updated if it can't be found. Or, if a story is marked as "closed"/"accepted", don't do anything to it. How would this be done? What would this look like in pipeline code? Not sure.

self explanatory

- put: tracker
  params:
    name: $NAME_FILE
    name_file: path/to/name/file
    description: $DESCRIPTION_FILE
    description_file: path/to/description/file

I love Concourse. But I can't safely run it without hosting my own server. Is there some way the CI/CD process can be done for free?

works just fine: DEMO. Upgrade to p-mysql 1.10.12 in Sandbox
can't find any matches: DEMO. Upgrade to p-mysql v1.10.12 in Sandbox

Whenever integers are used as param values, Concourse treats it as scientific notation. Whenever you run set-pipeline, every single time it treats it as a changed value. One option would be to accept large int values as strings && parse them as ints within the code.

why are there so many dockerhub pulls? bots? is that a fluke from long ago when this was used for development testing? ... but 7 million? I imagine a lot of it is just crawlers, but still, taking a screenshot to track, I suppose.

this entire project should really just be a task

number reading right before making the dockerhub repo private

{
"user": "aegershman",
"name": "pivotal-tracker-resource",
"namespace": "aegershman",
"repository_type": "image",
"status": 1,
"description": "",
"is_private": false,
"is_automated": false,
"can_edit": true,
"star_count": 0,
"pull_count": 7822490,
"last_updated": "2018-03-14T14:58:19.182832Z",
"is_migrated": false,
"has_starred": false,
"full_description": "",
"affiliation": "owner",
"permissions": {
"read": true,
"write": true,
"admin": true
}
}

Could the Dockerfile build flow be cleaner? Clearly there is something inefficient about the way things are currently. Although it is working, there's clearly room for improvement. I mostly hacked it together to get to an MVP, so it should be given more attention