Skip to content

TypeError instead of InvalidPointError on VerifyingKey.from_public_point(INFINITY) #341

New issue

Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? # to your account

Open
akarve opened this issue Jun 11, 2024 · 1 comment
Labels
bug unintended behaviour in ecdsa code help wanted
Milestone

Comments

@akarve
Copy link

akarve commented Jun 11, 2024

I was experimenting to understand how the library handles failures. I don't expect the following line to work but I do expect it to raise a different error (MalformedPoint, not Type).

VerifyingKey.from_public_point(INFINITY)
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
Cell In[3], line 1
----> 1 VerifyingKey.from_public_point(INFINITY)

File ~/miniconda3/envs/de/lib/python3.9/site-packages/ecdsa/keys.py:170, in VerifyingKey.from_public_point(cls, point, curve, hashfunc, validate_point)
    168 self.default_hashfunc = hashfunc
    169 try:
--> 170     self.pubkey = ecdsa.Public_key(
    171         curve.generator, point, validate_point
    172     )
    173 except ecdsa.InvalidPointError:
    174     raise MalformedPointError("Point does not lay on the curve")

File ~/miniconda3/envs/de/lib/python3.9/site-packages/ecdsa/ecdsa.py:151, in Public_key.__init__(self, generator, point, verify)
    149 n = generator.order()
    150 p = self.curve.p()
--> 151 if not (0 <= point.x() < p) or not (0 <= point.y() < p):
    152     raise InvalidPointError(
    153         "The public point has x or y out of range."
    154     )
    155 if verify and not self.curve.contains_point(point.x(), point.y()):

TypeError: '<=' not supported between instances of 'int' and 'NoneType'

Related: for cases where one wishes to avoid the point at infinity is this the right way to detect it if point == ellipticcurve.INFINITY? I kind of expected infinity to vary by the curve family but there's nothing under eg SECKP256k1 that I can find.

@tomato42
Copy link
Member

the INFINITY is a special kind of object that's it's a point at infinity for all curves and all representations, and yes, it should raise MalformedPointError here

@tomato42 tomato42 added the bug unintended behaviour in ecdsa code label Jun 12, 2024
@tomato42 tomato42 added this to the v0.20.0 milestone Jun 12, 2024
# for free to join this conversation on GitHub. Already have an account? # to comment
Labels
bug unintended behaviour in ecdsa code help wanted
Projects
None yet
Development

No branches or pull requests

2 participants