SPIKE: Investigate alternatives to stripes-connect
Description
Environment
Potential Workaround
Checklist
hideTestRail: Results
Activity
Zak Burke October 2, 2022 at 12:14 PM
We investigated swr and react-query. We chose react-query.
Cate Boerema September 7, 2020 at 12:52 PM
can you please give this a priority? Thanks!
Cate Boerema June 19, 2020 at 11:20 AM
can you please give this a priority? I think maybe it should be changed to tech debt, as well?
Cate Boerema November 19, 2019 at 4:02 PM
Assigning to Core Functional for now just to make sure it's visible in a backlog. says this could also go to Stripes Force.
Some comments from Zak in a separate thread which might be good context to keep here:
FYI I want to point out https://folio-org.atlassian.net/browse/STRIPES-663, “SPIKE: Investigate alternatives to stripes-connect” which came out of a long conversation about stripes-connect at #stripes-architecture last week. We acknowledged that there would be no activity before the end of the year but wanted to get it out there with the intent to do some real work in Q1/Q2 2020.
A big part of the conversation was whether looking at a REST-oriented alternative, i.e. not GraphQL, would be a less onerous conversion because it would be more like changing the brand of gas you buy for your car rather than changing from gas to diesel and therefore needing to retool the whole engine.
FWIW, this fell out of a discussion of https://folio-org.atlassian.net/browse/CHAL-91, where the conclusion was that it would vastly simpler to change how the MCL from stripes-components should ask for more data rather than trying to fix the bug in stripes-connect that causes duplicate requests. IOW, we weren’t just griping about stripes-connect; it presents real obstacles.
Zak Burke November 14, 2019 at 5:13 PMEdited
Details
Details
Assignee
Reporter
Zak BurkePriority
TBD
stripes-connectwas written as a proof-of-concept many years ago. It has been a workhorse for us, but we have never been able to dedicate the resources necessary to make it a fully-fledged library. Since it was first written, many other libraries have entered the arena that are fully-fledged, widely-used, battle-tested, well-supported and un-hyphenated (at least in the case of GraphQL). We have recently bumped up against some ofstripes-connect's rough edges in the form of bugs and missing features like caching. Before we look at spending some time to refactorstripes-connect, we think we should spend some time researching other solutions.