Try Rubi as a git submodule

[rocky]

Thoughts about doing this?
@aravindh-krishnamoorthy ?

[mmatera]

@rocky, I think this would be OK once mathics-ruby loads and works properly. By now, I would do the opposite: add a github-action test on mathics-ruby, and check against the branches of mathics-core that improve the required support.

[rocky]

@rocky, I think this would be OK once mathics-ruby loads and works properly. By now, I would do the opposite: add a github-action test on mathics-ruby, and check against the branches of mathics-core that improve the required support.

Yes, that makes sense. But here we need #1176 to finish. This corrects some bugs in handling the "optional" operator in pattern matching. When you get a chance, please look that over since you know far more about Mathics3's pattern-matching code (and WMA) than I. Right now, these changes are pretty small.

[aravindh-krishnamoorthy]

Indeed, this sounds good once basic Rubi starts working...

[rocky]

Now that Rubi has merged, this is your last chance to weigh in before we merge this.

Of course, in order to use Rubi, you'll have to use a branch to make TimeConstrained work. @aravindh-krishnamoorthy started a discussion thread for longer-term solutions.

[mmatera]

Ok, let's try this. LGTM