[Merged by Bors] - refactor(Algebra/Lie/BaseChange): use new tensor product machinery

[eric-wieser]

---

[ocfnash]

It's great that this works but given how slow it is, it seems like we're still not ready to use TensorProduct.AlgebraTensorModule.tensorTensorTensorComm here. How about just updating the comment and leaving the implementation as it was?

[eric-wieser]

I managed to speed it up, I didn't need the heterobasic curry after all, and that just slowed it down.
It's still a bit slower, but it's manageable.

[ocfnash]

Thanks, now that this is not too slow, I think it's worth doing this to have an application of TensorProduct.AlgebraTensorModule.tensorTensorTensorComm.

It looks like further golfing is now possible (we should no longer need both bracket_tmul' and bracket_tmul and even bracket_def should not be necessary) but I don't think it's worth spending more time on this.

bors merge

[eric-wieser]

Thanks, now that this is not too slow, I think it's worth doing this to have an application of TensorProduct.AlgebraTensorModule.tensorTensorTensorComm.

There's always #6306 if you want another one :)

[bors]

Pull request successfully merged into master.

Build succeeded!

The publicly hosted instance of bors-ng is deprecated and will go away soon.

If you want to self-host your own instance, instructions are here.
For more help, visit the forum.

If you want to switch to GitHub's built-in merge queue, visit their help page.

• Build
• Check all files imported
• Lint style