fable-compiler
Ideas to improve the REPL
-
Let the user save multiple samples with different names in browser storage
-
Let users easily share snippets like http://fssnip.net/ (requires a server)
-
Enable tutorials that go step by step and change the code in the REPL.
-
Make it easy to embed the repl in other sites (so people can make their own tutorials).
This should be already possible with an
iframe, and I also made it so the full repl won't be reloaded on hash changes (this could be a way to paste new code in the repl), but I guess we need a query parameter to disable some editors to fit the repl in a smaller size.
Let the user save multiple samples with different names in browser storage
Should be doable by adding a new section in the sidebar. We need to think about it because it isn't easy to show lot of info in the sidebar and perhaps we will need a modal to offer good UX to choose, clean the info in storage.
Let users easily share snippets like http://fssnip.net/ (requires a server)
This is probably doable even by using free ressource like Firebase functions. From what I understand firebase free plan is pretty generous.
Make it easy to embed the repl in other sites (so people can make their own tutorials).
I think for this one we should think in the other direction. If we create the REPL as an Elmish component then even the officiel "Fable repl" site would be a consumer of the REPL components.
I think we could offer several components like:
- an editor that's integrate with FCS for intellisense and so
- explain how to trigger a build
- a component to generate the iframe view
Like that people can pick what they need like the editor and the build trigger. But only display the output code etc.
Yes, using a component was the original idea, but given that setting up the infrastructure is not trivial (worker, loading assemblies), I thought maybe an iframe was easier for now. But you're right, we should probably strive to create the component and let users configure it as much as they want :+1:
Maybe some offline capabilities? Reloading takes too long and it downloads a lot of data. Having a service worker to cache the files might be a nice addition.
That's a great idea @nhowka! I've not much experience yet with service workers, do you have any pointers? Also, is it an issue that the files are downloaded from the web worker?
@alfonsogarciacaro I may be wrong, but I believe that the service worker can capture fetch events from everywhere. I'll make some tests and check if just creating a cache to save the requests on it is enough.
