yes, no errors 👍

I just upgraded an app based on fulcro-template 1.x to 2.x and am now getting 6 of the following errors:

[ 11.504s] [fulcro.client.impl.application] Mutation  fulcro.inspect.ui.transactions/add-tx  failed with exception Error: Assert failed: Path [:fulcro.inspect.ui.transactions/tx-list-id [:fulcro.inspect.core/app-id rimdb-ui.ui.root/Root] :fulcro.inspect.ui.transactions/tx-list] for append must target an app-state vector.
(vector? (get-in state data-path))

and that is using [fulcro-inspect "2.0.0-alpha5"]

seems to work fine though

Not sure what you’re asking. There is a github fulcro-template that you can clone (with history). The problem there is providing options. I’ve decided to lean more towards treating templates like a pure starting point. Once you’ve type “new”, it’s yours to evolve however you see fit.

Sure, but is that good for your users? In a project that’s growing as fast as fulcro, including the growth of shadow-cljs, it doesn’t scale when all your users need to manually maintain their boilerplate and add deps every time a new feature lands.

Feel free to suggest a specific solution, code it up, and send a PR. I’m currently doing what I can with the time I have. I don’t see how “updating a version” or “tweaking a config” doesn’t scale for users any worse than “fixing merge conflicts” from changes you didn’t make.

I really don’t think there is a great solution…things change, and I don’t see a way to make it substantially easier. I’ve tried it all of the ways I can think of, and over time, it just doesn’t seem to matter.

I guess perhaps I don’t know what you mean by a “library”. Do you mean like I’ve done with fulcro-template (the non-lein template)?

Will do. I’m working on a project to show the difference. What I mean by library, is something you don’t generate or clone but use (add it to your build.boot / project.clj). The author can change a library to improve all users’ experience without requiring them to take any action. This way long lived projects become updatable instead of a copy & paste mess. Anyway @tony.kay please don’t take this personally and thanks for the effort you put into clj/s tooling and fulco

Not taking anything personally 🙂 Just trying to see your proposed solution.

Requirements include: dev, runtime, uberjar generation, CI testing (client and server), dev-time use of fulcro-spec in browsers for client and server. I guess you could sorta make lein/boot plugins in a library, but that seems painful to maintain and get working.

oh, and devcards support

and most users will need to customize a lot of that based on their needs. If you’ve got the cycles to figure out something better, happy to see it and promote it!

I think all you need is the repl. I now have a prototype that installs node-modules, shadow-cljs for dev, and index.html file, starts an http server with custom routes and much more after just calling a simple dev-start function from my library. It’s easy to see how this could be extended to provide CI files, and even augment the project.clj, deps.edn files if necessary. A config must be passed in by the user, on a breaking version change, the library can explain exactly which keyvals are wrong/missing, etc. Currently just building a prototype for myself. If it pays of, I’ll share my findings here.

Hey, sounds great!