New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Don't exclude galaxy.yml from collection when doing build #68383
Comments
Files identified in the description: If these files are inaccurate, please update the |
I think this is a good idea we can continue to create the The benefit of using just
|
That could enable #61680 |
I'm not totallly opposed to this... My primary concern was dealing with the dual files of galaxy.yml/files.json (which one take precedence when both exist?), and having to load and ignore all the file metadata at runtime from files.json... But if we're willing to rally around galaxy.yml for all runtime needs (which means we have to start preserving it in thepublished artifacts), it's less of an issue, and solves some other issues as well. So let's talk. |
My thought around this was to ignore The only sticky bit here would be verify and whether we eventually add something like verification at runtime. Using |
So i would:
Putting everything in one file does not scale well in the end and creates a very complex format. While having simple specific files for certain features seems easier for content authors in the end. Another note, right now we don't really read the contents of galaxy.yml except to build manifest and we don't really read manifest except at install time for dependencies. At runtime we use 'both files' to figure out if something is a collection but only by existence, not by reading contents. |
@bcoca - Yes and no... I'm already getting confused about where certain metadata about my collection should go. Part of that is the fact that things are not all standardized yet, but part of it is that it seems like there are currently already a couple locations for metadata, and (as mentioned in the original comment on this issue) it looks like there eventually may be 3 or 4 (or more) files to stash metadata inside. |
Some of that was also sort of by design, eg: 99% of people won't know or care what So it's a bit of "damned if you do..." IMO. I can live with it either way if there's a lot of support for Bigger picture: we need to get this resolved pretty soon... Maybe something to bring up at this weekend's EU contrib summit @gundalow ? |
Hasn't been any movement here from the PM side of things, so going to just close for now. |
Instead of ignoring
galaxy.yml
when building collections, let's embrace it!SUMMARY
Currently there are a few new features being baked into collections like:
meta/action_groups.yml
meta/main.yml
meta/runtime.yml
meta/routing.yml
The general problem I have is we're starting to see an explosion of metadata files, and not only will this be hard to document and find a good pattern for in ansible/ansible, it will be hard for collection maintainers to remember all the various places different types of collection metadata should go.
In discussing this with @jborean93 on IRC, he mentioned one possible solution is to ship the
galaxy.yml
with the collection, and use that file as the home for all collection metadata. It already has like 99% of it, so why not store everything in one place?There's precedent: Node.js/NPM
package.json
, PHP/Composercomposer.json
, Ruby/BundlerGemfile
...ISSUE TYPE
COMPONENT NAME
Collections
ADDITIONAL INFORMATION
Related: https://github.com/ansible-collections/kubernetes/issues/58
The text was updated successfully, but these errors were encountered: