Drupal Association members fund grants that make connections all over the world.
This project has had many incarnations, first was the http://drupal.org/project/plugin_manager project in contrib, then and the related issues. This push got us almost to the point of completion before Adrian wrote up a review of it at http://groups.drupal.org/node/24709, which caused some of the authors and contributors to rethink the architecture.
Dries, myself and Adrian met and decided to provide a very limited spec which:
- Is possible to finish by D7 code freeze
- Provides a minimally viable product
- Can be disabled by people who do not want it active on their site(s)
We've created this new issue and its sub-issues to have a fresh start since the previous issues may at this point be confusing, and are incredibly long and hard to follow.
What is outlined here is final and has Dries's approval. This is not to say you can't get suggest stuff you want to see differently implemented, but it is up to Adrian and myself to take a call on it, and we will not increase the scope. This is not because we want to be jerks or power mongers, it's because we want this done by September 1st. At DrupalCon, there will be a BoF where we can scheme phase 2 stuff. Cool?
What we are trying to do (by sep 1st)
Allow Drupal administrators to install new modules on their site and update existing modules without knowing how to use an FTP client or SSH. This will open up the user base of Drupal and make evaluation easier. It is also hoped that will lead to greater stability in contributed modules as they get wider non-developer adoption.
What we are not trying to do (by sep 1st)
(but would love to do in the future)
- Multi-site installs upgrades
- Multiple repository sources
- Version control integration
- Circular dependencies
- Install profile integration
- Backup and recovery
- ... add your feature here
Of course, the more work we all do, the more we get done.
The Backlog and User stories
This project consists of 2 Epics (overarching goals) and a series of stories which fulfill those Epics. For each story, there will be one or many issues which fulfill it. For us to reach a minimally viable product, all of our Must Have (P1) stories must be complete.
Epics for D7 Code Freeze
As a Drupal admin, I should be able to keep my Drupal contributed modules and themes up to date without knowing how to use an FTP client or SSH.
As a Drupal admin, I should be able to browse the available modules and themes on Drupal.org, and paste a URL to a release archive or upload a release archive onto my site and have it be uploaded to the module or theme directory (but not enabled).
Stories and related issues to complete before D7 code freeze
Users using PM should not have PM crash if a contrib module is faulty after an upgrade / install so that there is a chance of fixing it from the GUI
As an Admin, I can upload a tarball containing a plugin, or specify a url to a package and install it
As an Admin, I want to update my modules and themes from my browser so that I don't need to use FTP or SSH directly
What started out as an attempt to make small patches ended up in one honking patch:
Because we can't get anything committed 1/2 way, it was kinda difficult to split this one up. That being said, it seems to work pretty well, so we just need reviews to get it in.
Follow up issues: