Should probably drop/update angular-pouchdb dependency

[nolanlawson]

angular-pouchdb is a bit outdated, and it actually depends upon PouchDB 2.1, which is unfortunate because we had a Bower bug back then where we uploaded the entire Selenium JAR to Bower. Which means it takes forever to download:

Maybe you can just fork angular-pouchdb instead?

[danielzen]

I originally used https://github.com/angular-pouchdb/angular-pouchdb and had to modify it to serve my purposes. However, I switched to https://github.com/wspringer/angular-pouchdb because

a) It supported what I needed out of the box
b) I thought it was more accepted.

I will likely fork from the tlvince version. Thanks for the advice. That's the thing about just using someone else's packaged module: You don't always check to see how well it was written. At first I was just happy it worked!

[nolanlawson]

Yeah, there's unfortunately been a lot of fragmentation in the "Angular + PouchDB" space. Was hoping something could emerge as the definitive solution, in the same way that (I think) I succeeded with ember-pouch. ng-pouchdb looks really promising for that. :)

[johannesjo]

Any news on this one?

[tlvince]

👍 in forming a canonical "angular-pouchdb".

I've been slowly maintaining https://github.com/angular-pouchdb/angular-pouchdb and would of course love to pull people over to it, but happy to go with whatever works best.

[nolanlawson]

What I would really like to see in an Angular+PouchDB plugin is a nice set of PouchDB-powered directives, so that I can write pure HTML and have that data automatically bind to a PouchDB. Wrapping PouchDB's promises and applying $rootScope.$apply afterwards is okay, but for performance it would be much better if all that logic was contained within a directive, because then you could $apply the affected element, instead of the entire page. Also, the user wouldn't even have to write any JavaScript.

Also, many of these plugins seem light on documentation and examples. I've always been a big proponent of generous documentation, and I think in the Angular space this is even more important, because Angular developers tend to skew more towards newbies.

TBH in the present situation, I've been recommending Angular developers to just use straight-up PouchDB and avoid any of the plugins. I hope this situation changes soon! 😃

[tlvince]

Wrapping PouchDB's promises is a convenience and has its uses in services, but you're right; these projects could add more value via directives 👍

[tlvince]

NB, ng-pouchdb's 3(/4?)-way binding looks pretty neat!

[danielzen]

I really want to work more on this. I've just been under the gun on client work. Everytime I talk on the subject, I add to the repo. Unfortunately, I haven't gotten any of my recent talk proposals on this subject approved :(

[danielzen]

@tlvince I'm planning on moving back to your version of angular-pouchdb, but you never approved my pull request which had needed functionality, so I initially used a forked version.

Also, had some issues with including the pouchdb module "PouchDB" vs "pouchdb" when using your module, and instantiation:

var database = new PouchDB(reference);
vs
var database = pouchdb.create(reference);

seems like your version vs wspringer's are different in these regards.