Invent compatibility

[mhsmith]

This is a meta-issue listing the items that would be required for Invent to base itself on Toga, along the lines of this diagram (source):

The following is mostly based on this comparison chart of the two projects' existing APIs.

General concepts:

• Style property redesign #3011
• Implement ValueSource #3012
• Event design differences – channels vs handlers (no issue yet)

Flexbox enhancements:

• Improve alignment options #1194
• Equivalent of CSS gap (padding between children) #1943
• Add Row and Column subclasses of Box #3010

Other wIdget enhancements:

• Button purpose property #3013
• Border or frame support #849

Testing:

• Add testbed support for web backend #2662

---

New widgets – these would NOT be blockers, as long as the architecture allows them to be supported on Invent only at first:

• Grid layout #2162
• Add widgets to differentiate between Switch and Checkbox #2225

[HalfWhitt]

The comparison chart doesn't load for me.
Could you upload and inline your images so they're in the post?

[mhsmith]

The comparison chart doesn't load for me.

Sorry, I've fixed the link.

Could you upload and inline your images so they're in the post?

Done.

[HalfWhitt]

Assuming all this works out... would Toga still even need a separate web backend? From that diagram, it looks like as long as the "Toga app components" are also implemented in toga_invent, a fully Toga app could then use that to display in a web browser. At which point it would make more sense to name the shared backend toga_pyscript.

[freakboy3742]

@HalfWhitt It depends a little on the final form of rendering. Any web-based implementation will likely share a lot of common infrastructure and tooling; but there might be some differences depending on exactly how styles and web components are used.

Toga-web and Invent both use pyscript as a bridge; but when rendering widgets, Invent tries to keep things as close to "pure" HTML + CSS as possible, and uses paperCSS. Toga-web has historically used Bootstrap, and currently uses Shoelace (soon to be renamed "Web awesome") to provide a richer set of widgets. This doesn't matter as much for simple widgets like Button, but might for more complex widgets like "Tree" that doesn't have a "native" HTML widget that can be styled.

So - what is labeled as "toga-invent" in that diagram might end up being;

• The same "toga-web" we currently have, but with Toga switching to using PaperCSS
• The normal "toga-web" but with a "paperCSS rendering plugin" provided by Invent, and Toga providing a "shoelace rendering plugin" by default.
• A new "toga-invent" backend which builds on a base "toga-web" backend that can't be used on its own. "toga-invent" would add paperCSS-based rendering; with Toga would then ship a toga-shoelace backend for shoelace-based rendering (or use shoelace by default in toga-web)

We need to do a little more investigation before it's going to be clear which approach will be most viable.