Published a couple more blog posts to https://chrisodonnell.dev about setting up a fulcro server built around a pathom parser. Happy to hear any feedback. :)

The develop branch of the rad-demo has been updated. I refined the following: • Forms now have a way to be marked read-only, and the rendering of that is nicer • Reports have two ways of turning columns into links: a convenient form-link, and a more general ::report/link that is just a fn • Passing parameters as an extra parameter to route-to! will push those into the report parameters of a report on route. See Row actions on invoice report that links to a specific account’s invoices. • Routes of reports include their parameters, making the report completely bookmarkable • Reports now have report-level and row-level actions. • Fixed HTML5 fwd/back nav for route denials. I.e. if you have an unsaved form it can block fwd/back • Generalized application routing to route-to! in rad routing ns. This allows you to route to something without needing to know it’s UI path, even though it might be deeply nested. (limitation: there must only be one of that kind of component in your query tree) • Read-only fields now use Semantic UI’s transparent class, which just hides the border, keeping the input (and its formatting) in tact. • renamed a few keywords (sorry, still making things consistent). NOTE: The docs (RAD book in particular) are out of date. Consider the source of the demo the correct source for the time being.

NOTE 2: Use the Datomic db driver for now. The SQL one mostly works, but things like enum save are broken (https://github.com/fulcrologic/fulcro-rad-sql/issues/7). See other open issues on fulcro-rad-sql

I tried the fulcro-inspect electron version today. The alternative "DB Explorer" is nice, but I noticed the "Index Explorer" is very barren and unlike the one I used via the Chrome extension. It looks like no CSS was pulled in? Or is it a different version of the Explorer?

@tony.kay I know that the SQL adapter in RAD should not be used but I need it 🙂 Just wanted to let you know that while (go)works fine, after a (restart) I get a strange Pathom error:

Execution error - invalid arguments to com.wsscode.pathom.connect/register at (core.clj:2353).
core.clj:2353

-- Spec failed --------------------

Function arguments

  (...
   [... [... ... ... [[:com.wsscode.pathom.connect/sym com.example.model.account/login] ... ...] ... ... ...]])
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

should satisfy

  map?

or value
...

Probably you messed up one of your resolver definitions

which spec failed?

3.1.23-SNAPSHOT (just pushed to clojars) has a few new things on it I’m still testing/refining: 1. No longer uses cljsjs at all. You MUST use npm to get react/react-dom, and if you’re using std cljs pull those in via webpack (https://clojurescript.org/guides/webpack), or if you’re using shadow-cljs you’re already fine. 2. (1) allows you to now try preact instead of React! I know of one bug at the moment: preact doesn’t like it if you give it js vectors, so any map over children currently has to be wrapped in a clj->js . I can probably fix that, but want to make sure I don’t screw up performance for React users. 3. Hook support now has configure-hooks-component! that allows you to make a hook component without defsc (even dynamically at runtime, like configure-component!). 4. IGNORE THIS ONE…MIGHT CHANGE: There is now a ui-input in the dom namespace. I’m trying to manage React’s tendency to jump the cursor to the end of inputs. Newer versions of react have more async, and it is resurfacing at times. That control is like input, but onChange and onBlur receive just a string instead of an event. It seems to work better in scenarios I can reproduce, but still isn’t perfect. NOTE: preact does not seem to misbehave like React on this one.