Closed artfulrobot closed 7 years ago
@artfulrobot, this is super! With the year starting to dwindle down, this API v3 version is gaining importance.
Can you provide a status update on where this effort is today? What needs to be done yet?
Very interested as well. Does this module need more support to continue?
@zkrebs support is always welcome, we invest hundreds of hours per year into the extension without any funding so it doesn't always get priority treatment. If you are able to invest then it would allow us to push the changes through with some testing etc. If you'd be interested in sponsoring some of the work please do let us know at parvez@vedaconsulting.co.uk
I will send this to my orgs team and see if there is any budget for us. We are a very small non profit using this extension, it is great, and we want to find a way to contribute to the projects we're using.
@ddoligalski Hi. Sorry for slow reply, been on holls :-) I did the work on the api3 version which is a massive rewrite.
The api3 version that I wrote and submitted to Veda is in production use on my own sites and seems to be working well. I'm not sure what veda are doing in terms of releasing it to all as the main version? Basically the important thing is that on install this will make some backwards-incompatible changes to your config, so a downgrade will fail (always take full db backup before upgrading anything...).
There is also a PR that builds on top of my work and adds multiple account support, however that broke a lot of the the tests. It is my opinion that software like this must have tests and that no code should be accepted until all tests pass; this ensures a stable platform. The contributor seemed to be working on those tests, but I've not heard from them in a couple of weeks.
@artfulrobot's work is out as v2.x release.
...And API v3 is miles apart. AND we currently store Mailchimp 'grouping' Ids that simply don't exist in v3; there needs to be a migration before the old API is removed.
I am about to release a PR that is pretty much a rewrite and uses the v3 API. It also introduces some changes, details coming soon, but I hope it will solve several other issues, too, notably performance e.g. it now pull-syncs ~5000 contacts in under a minute compared to ~15 minutes for the same dataset before. The matching up of contacts is hopefully a lot more successful now, too.
I've written dozens and dozens of tests for the code, too, to check it's working as it should, but also to document expectations for those weird cases. I've included a changelog and a technical document to explain the code.
I hope it should fix a few of the issues on this issue queue, too, so I'll reference them in the PR.
I though it would take a week. It's taken weeks! Watch this space...