bug: Invalid Jubjub curve generator

[sug0]

The curve generator of Jubjub is configured improperly, see the following snippet:

gnark-crypto/ecc/bls12-381/twistededwards/curve.go

Lines 49 to 50 in b81524c

	curveParams.Base.X.SetString("23426137002068529236790192115758361610982344002369094106619281483467893291614")
	curveParams.Base.Y.SetString("39325435222430376843701388596190331198052476467368316772266670064146548432123")

It follows from the reference at https://std.neuromancer.sk/other/JubJub that:

Gx := 0x11dafe5d23e1218086a365b99fbf3d3be72f6afd7d1f72623e6b071492d1122b = 8076246640662884909881801758704306714034609987455869804520522091855516602923
Gy := 0x1d523cf1ddab1a1793132e78c866c0c33e26ba5cc220fed7cc3f870e59d292aa = 13262374693698910701929044844600465831413122818447359594527400194675274060458

[ivokub]

Indeed, I cannot find any reference to the base point value we're using. Seems @gbotrel implemented it, perhaps you recall?

[gbotrel]

mhhh git blame pointed to me? probably a refactor -- generator choice would probably be @yelhousni :)

[sug0]

I think this generator was first introduced in gnark-crypto in 771aa8c

but it is in fact already present in the initial commit of gnark, see: https://github.com/Consensys/gnark/blob/1e90a34c9596ba60bc38ead2190836e9e397686a/cs/std/reference/algebra/twisted_edwards/edwards_bls381.go#L39-L40

who knows where it originated from :-)