This change adds text around how plugin selection for plugins defined in
a require_plugins block work. It also adds a small call out to users
that third party plugins are not verified by HashiCorp.
* document builder id of artifact for each builder's output on website
* document postprocessor artifact ids on website
* Fix links to new communicator page
Co-authored-by: Wilken Rivera <dev@wilkenrivera.com>
* Packer has no state like Terraform has
* init command docs: tell where a plugin will be installed
* docs: add a note that packer init does not work with legacy JSON templates
* docs: add a not that packer init does not install single-plugin binaries
* the plugin getter => packer init
* Update init.mdx
* grammar
* link to how to install plugins manually
* Update website/content/docs/commands/init.mdx
Co-authored-by: Sylvia Moss <moss@hashicorp.com>
* Update website/content/docs/commands/init.mdx
Co-authored-by: Sylvia Moss <moss@hashicorp.com>
Co-authored-by: Sylvia Moss <moss@hashicorp.com>
This adds the new `required_plugins` block to be nested under the packer block.
Example:
```hcl
packer {
required_plugins {
aws = {
version = ">= 2.7.0"
source = "azr/aws"
}
azure = ">= 2.7.0"
}
}
```
For example on darwin_amd64 Packer will install those under :
* "${PACKER_HOME_DIR}/plugin/github.com/azr/amazon/packer-plugin-amazon_2.7.0_x5.0_darwin_amd64"
* "${PACKER_HOME_DIR}/plugin/github.com/hashicorp/azure/packer-plugin-azure_2.7.0_x5.0_darwin_amd64_x5"
+ docs
+ tests
The docs for top-level and build-level source blocks did not make it clear enough, early enough in the page, how packer merges their contents and fails when ambiguities arise from redefinition. I know the info it there, but it's at the end and I feel that my modification shortens the page overall while putting the most useful info front and center.
Also, there is simply an error in the note at the end of this page, which further confuses the situation for new readers trying to use HCL and reuse source blocks. It's referring to a non-existent amazon source which is probably a past copy/paste error.
It took me a day to get re-usable source blocks working, bc I skimmed the page at first (I know slap my wrist), but didn't really grasp what it was saying in the Note at the end, and found it confusing. It wasn't until I found the Issues and PRs related to this feature getting added last April/May that I realized how it was implemented that I started to grasp what this was saying.
* Allow locals to be delcared as individual blocks, and give them the Sensitive flag
* add docs for new local block
* linting
* add tests
* modified parsing to use schema, check for dupes properly
* update comment
fix wording a liiitle
* add tests for duplicate variables definition in two different files
* remove unnecessary slice initialisation
* fix crash by returning when decode error is hit
* parseLocalVariables: only treat a local vars if its not nil
also return in case of error
return locals in case of error too
* fix duplicate_locals test for windows
Co-authored-by: Adrien Delorme <azr@users.noreply.github.com>
* Test against deployment url
* Remove infinite redirect configuration
* Add DEPLOYMENT_URL for builds against master
* website: Update README
* tip on seconds
* Test with GHA timeout_minutes clause
* Add continue on error for poll job
* Add empty url check
* Move to pull-request path filter
* Remove www for packer.io
* Apply suggestions from code review
* Update path filter
Vercel deploys on any change under the website directory.
* Use custom action as test
Co-authored-by: Adrien Delorme <azr@users.noreply.github.com>
