Feat/add exp gate

[ainta]

• added ExponentiationGate to common_data_for_recursion
• ExponentiationGate is needed for ed25519pod in from introduction-pods(expose some interfaces for external usage (from introduction-pods) #256)

[ed255]

As discussed in the Telegram chat, the idea for introduction gadgets is that they can use their own configuration and set of gates without requiring modification of the standard set of gates from the recursive MainPod. This is achieved by adding a wrapping circuit that "normalizes" the proof, see the current implementation of 0xPARC/introduction-pods#1 for reference.

I've found the source of this gate:

The recursive verifier calls this function https://github.com/0xPolygonZero/plonky2/blob/82791c4809d6275682c34b926390ecdbdc2a5297/plonky2/src/gadgets/arithmetic.rs#L288-L291
And I believe the length of exponent_bits is related to the proof size (number of gates).
For example I've tried generating a verifier from the Recursive Main Pod and the length is 19, which is below
self.num_base_arithmetic_ops_per_gate()=20.
In the ed25519 circuit this length is 22 which causes the verifier to use the ExponentiationGate.

So this gate is not related to a gate that the ed25519 circuit uses but the fact that the proof is very big and requires an extra gate to be verified.

If this was about a special configuration/common that the introductory gadget needs I don't think we should add this gate to the standard list, but since this is a generic case of a big proof, and currently the proof we're verifying in the Recursive MainPod is alreay big-ish, I think it's OK to add the ExponentiationGate to the standard list of gates.

Note that the situation would be different in approach (b) from #252 (but we can cross that bridge when we come to it)

[ed255]

Also could you rebase/merge with main?

[ed255]

LGTM!