Hello everyone,
I'm not sure who follows this list and who doesn't, but I wanted to announce to folks that I've now taken over as maintainer of Dispatch from Matthew de Detrich who took over briefly from Nathan.
I've used Dispatch to build a number of things over the years. I first used it in 2011, back in the dispatch-classic days. Since then it's been my regular go-to when I needed a good HTTP DSL in Scala. I've used it for a number of projects since then and it has been an indispensable tool on a number of occasions. I've only attempted
a few small contributions to date where Dispatch intersected with Lift, a project I also commit code for, but I'm excited about the opportunity to steward a library that has helped me out so many times.
TL;DR, don't expect very much to change with Dispatch. Hopefully, much of what you'll notice is that there's a bit more responsiveness to issues, pull requests, and on this mailing list.
There are a few things that are on my radar, however:
- First, given that Dispatch API has been as stable for as long as it has, I'm strongly considering the next release being marked as 1.0.0 and adopting the following version semantics afterward:
- Changes to the major version (leftmost number) would mean major API changes (particularly, entire methods changing meaning or disappearing), or enough internal changes that migration could be painful.
- Changes to the minor version (middle number) would mean new features and minor API changes (new methods, changes to arguments of methods, deprecations, etc). Migration for most cases is very straightforward.
- Changes to the patch version (rightmost number) would be bug fix only releases.
- Second, formulating and publishing some kind of release schedule and version support statement. It's unreasonable to think that I'll be able to maintain every active version of Dispatch that's ever been released. It's also unreasonable to think that folks are going to rush to update their app every time changes get pushed to Dispatch. I'm hoping to communicate around this clearly so that people using dispatch know what to expect.
- Third, dependency bumps to the existing Dispatch library. Since this will require some breaking changes under the hood, I don't think it's going to be possible to do this in the 0.12.x series and feel good about myself, but I'm currently thinking through how big, exactly, our changes should get before the 1.0.0 mark.
- Fourth, paring down the issue and pr backlog to immediately actionable items.
Once the above are in place, my hope is that maintaining Dispatch will get to a point where it's a matter of small, regular executions instead of heroic efforts.
With that said, please let me know if you have any feedback about the above or anything related to Dispatch. Messages on this ML will now appear in my inbox, so I'm looking forward to hearing from you.
Cheers,
Matt