Closed jogibear9988 closed 7 years ago
@bergie we have a circular dependency here then. Not really, we just have to bite the bullet at the same time.
Proposed way forward is to:
Longer term we're probably getting rid of Polymer completely, but this we will do component/widget by widget as they get other improvements - not rewriting for the sake of rewriting. Also keeping thin Polymer wrappers, and then at some point when the innards are all Polymer free we can drop Polymer in one go.
Old version of Polymer is last sticking point in our VersionEye checks as well:
Dependencies with unknown licenses: 1
https://www.versioneye.com/pullrequests/57e13210c2cd3f0018cff57c
Migration guide: https://www.polymer-project.org/1.0/docs/migration
Polymer 2.x is now in preview as well: https://www.polymer-project.org/2.0/docs/about_20
Unfortunately it seems that Polymer's NPM packaging is somewhat lacking... Polymer/polymer#326
However, we have noflo-polymer working with Polymer 1.x from NPM now, so that is a start noflo/noflo-polymer#16
Polymer 2.x says explicitly to not use it for production, so don't see a reason to jump onto that at this time.
(and the migration path they document for Polymer 2.x is via Polymer 1.x)
Some migration notes from the-graph experience at https://github.com/kristianmandrup/the-graph/blob/polymer2/Polymer2-migration-notes.md The elements in that branch have been ported to Polymer 2.x, ref https://github.com/flowhub/the-graph/issues/345
Depends on the-grid/the-graph#222