Usage of newer ecsda package

[gstarovo]

fixes #373

This change is 

[tomato42]

Reviewed 3 of 7 files at r1, all commit messages.
Reviewable status: 3 of 7 files reviewed, 9 unresolved discussions (waiting on @codeclimate[bot] and @gstarovo)

---

.github/workflows/ci.yml line 8 at r1 (raw file):

      - master
      - tlslite-ng-0.7
      - ecc_changes

this will need to be removed from the final PR

---

tlslite/keyexchange.py line 751 at r1 (raw file):

        key = kex.get_random_private_key()
        if isinstance(key, ecdsa.keys.SigningKey):
            ecdhXc = bytesToNumber(key.to_string())

if it's a SigningKey we shouldn't change it to numbers, the ECDHKeyExchange needs to accept inputs that are both SigningKey and a number (so that it works with tlsfuzzer), but the rest of tlslite-ng should expect that the private keys are the SigningKey object, not an int

---

tlslite/keyexchange.py line 1031 at r1 (raw file):

            writer = Writer()
            writer.add(4, 1)
            writer.bytes += point.to_bytes()

point.to_bytes('uncompressed') will do the proper format and won't require using a Writer()

---

unit_tests/test_tlslite_keyexchange.py line 1959 at r1 (raw file):

                                        abstractPoint[1])
        point = generator * cln_Xc
        

nit: whitespace

[tomato42]

Reviewed 3 of 7 files at r1, 2 of 3 files at r3, 1 of 1 files at r4, all commit messages.
Reviewable status: all files reviewed, 12 unresolved discussions (waiting on @gstarovo)

---

.github/workflows/ci.yml line 8 at r1 (raw file):

Previously, tomato42 (Hubert Kario) wrote…

this will need to be removed from the final PR

I was referring to the ecc_changes line, not to the whole branches section

---

tlslite/keyexchange.py line 25 at r4 (raw file):

        X448_ORDER_SIZE
from .utils.compat import int_types
from .utils.codec import DecodeError, Writer

Writer is unused now, isn't it?

---

tlslite/keyexchange.py line 1051 at r4 (raw file):

        try:
            abstractPoint = ecdsa.ellipticcurve.AbstractPoint()
            point = abstractPoint.from_bytes(curve.curve, peer_share)

no, we should use VerifyingKey.from_string() here, or better yet ecdsa.ecdh.ECDH.load_received_public_key_bytes()

---

tlslite/keyexchange.py line 1057 at r4 (raw file):

            raise TLSIllegalParameterException("Invalid ECC point")

        S = ecdhYc * private

and if we use ECDH class, we can use generate_sharedsecret_bytes()

---

tlslite/tlsconnection.py line 1186 at r2 (raw file):

            private = bytesToNumber(key.to_string())
        else:
            private = key

why we need those changes?

[tomato42]

Reviewed 3 of 3 files at r6, all commit messages.
Reviewable status: all files reviewed, 11 unresolved discussions (waiting on @gstarovo)

---

tlslite/keyexchange.py line 1027 at r6 (raw file):

        """Calculate public value for given private key."""
        if isinstance(private, ecdsa.keys.SigningKey):
            private = bytesToNumber(private.to_string())

why not return private.verifying_key.to_string('uncompressed')?

---

tlslite/keyexchange.py line 1055 at r6 (raw file):

            ecdhYc = ecdsa.ellipticcurve.Point(
                curve.curve, point[0], point[1])
            

nit: whitespace

---

tlslite/keyexchange.py line 1063 at r6 (raw file):

            return bytearray(ecdh.generate_sharedsecret_bytes())
        if isinstance(private, ecdsa.keys.SigningKey):
            private = bytesToNumber(private.to_string())

the above two lines will never execute...

[tomato42]

Reviewed 1 of 1 files at r7, all commit messages.
Reviewable status: all files reviewed, 6 unresolved discussions (waiting on @gstarovo)

[tomato42]

looks good, thanks!