Adds to DER format of ECDSA signature with extended ECDSA-Sig-Value and ECDSA-Full-R by gstarovo · Pull Request #369 · tlsfuzzer/python-ecdsa

gstarovo · 2025-10-21T14:35:15Z

The changes expand encoding/decoding of the ECDSA signature in DER format: expanded ECDSA-Sig-Value format and ECDSA-Full-R.
The definition of the formats can be found in document section C.5, page 114.

The changes bring supporting functions like encoding and decoding of boolean according to ASN.1 standard.
Beside that, the changes add parameter and slightly modifies ECDSA signing function for enabling return of the whole point (instead of 'r').
The testing coverage of encoding/decoding with new function is also present.

Also, there are some formatting changes made with black formatter.

src/ecdsa/test_pyecdsa.py

gstarovo · 2025-10-21T15:41:22Z

I will provide more coverage for the newly added code, as the results of the checks indicate coverage decline.

src/ecdsa/test_malformed_sigs.py

src/ecdsa/der.py

tomato42 · 2025-11-03T16:07:33Z

src/ecdsa/der.py

+    body = string[1 + lengthlength : 1 + lengthlength + length]
+    rest = string[1 + lengthlength + length :]
+    if not body:
+        raise UnexpectedDER("Empty object identifier")


not an object identifier

changed to "Empty BOOLEAN value"

tomato42 · 2025-11-03T16:10:35Z

src/ecdsa/der.py

+    if body == b"\x00":
+        return False, rest
+    # the workaround due to instrumental, that
+    # saves the binary data as UF-8 string


this method shouldn't be passed Unicode strings ever...

added check to control if the body type is str, than it will be encoded to bytes

if isinstance(body, text_type): body = body.encode("utf-8")

tomato42 · 2025-11-03T16:11:44Z

src/ecdsa/ecdsa.py

+            point = ellipticcurve.AbstractPoint.from_bytes(
+                self.generator.curve(), r
+            )
+            r = point[0] % n


why this change though?

When encoding is in format of r in the siganture is ECPoint (ECPoint ::= OCTET STRING) to get the r, I firstly have to get the point from bytes and then extract the x and provide the same operation that is in the sign:
r = p1.x() % n

tomato42 · 2025-11-03T16:13:00Z

src/ecdsa/keys.py

+        :param accelerate: an indicator for ECDSA sign operation to return
+            an ECPoint instead of a number of "r" parameter.
+            Applicable only for ECDSA key.
+        :type accelerate: boolean


but we already are using the sigencode for specifying the output format...

Well, that is not nice workaround that is being propagated to ecdsa.py PrivateKey.sign() to signalize what type of the r (if int or a ECPoint) I want to return. So after that it can directly go to the sigencode function in correct format. I did not find a way how to do it differently, or how to get the whole point out of r directly in the sigencode function

Instead of r being just an int or ECPoint, maybe to create a new class, that will have both values all the time, so the sigencode will get all the values. I will show that change in a new commit. I will then squash it to one, if it will be accepted.

tomato42 · 2025-11-03T16:13:33Z

src/ecdsa/keys.py

+        :param accelerate: an indicator for ECDSA sign operation to return
+            an ECPoint instead of a number of "r" parameter.
+            Applicable only for ECDSA key.
+        :type accelerate: boolean


same here, sigencode should specify the format we want...

tomato42 · 2025-11-03T16:13:40Z

src/ecdsa/keys.py

            leak the key. Caller should try a better entropy source, retry with
            different ``k``, or use the
            :func:`~SigningKey.sign_deterministic` in such case.
+        :param accelerate: an indicator for ECDSA sign operation to return


src/ecdsa/test_der.py

…nd ECDSA-Full-R.

github-advanced-security bot found potential problems Oct 21, 2025

View reviewed changes

src/ecdsa/test_pyecdsa.py Fixed Show fixed Hide fixed

gstarovo force-pushed the ecdsa-sig-extend branch 5 times, most recently from 0961801 to e17d333 Compare October 23, 2025 09:39

github-advanced-security bot found potential problems Oct 23, 2025

View reviewed changes

src/ecdsa/test_malformed_sigs.py Fixed Show fixed Hide fixed

gstarovo force-pushed the ecdsa-sig-extend branch from e17d333 to 7f2a47b Compare October 23, 2025 11:09

github-advanced-security bot found potential problems Oct 23, 2025

View reviewed changes

src/ecdsa/test_malformed_sigs.py Fixed Show fixed Hide fixed

gstarovo force-pushed the ecdsa-sig-extend branch 2 times, most recently from 19ac7b2 to f173f56 Compare October 27, 2025 11:58

github-advanced-security bot found potential problems Oct 27, 2025

View reviewed changes

src/ecdsa/test_malformed_sigs.py Fixed Show fixed Hide fixed

src/ecdsa/test_malformed_sigs.py Fixed Show fixed Hide fixed

gstarovo force-pushed the ecdsa-sig-extend branch 5 times, most recently from 9c23305 to b046e1f Compare October 29, 2025 09:09

gstarovo requested a review from tomato42 October 29, 2025 09:53

tomato42 requested changes Nov 3, 2025

View reviewed changes

gstarovo force-pushed the ecdsa-sig-extend branch 5 times, most recently from 841cce2 to 9f5252e Compare March 26, 2026 10:40

Adds to DER format of ECDSA signature with extended ECDSA-Sig-Value a…

4984db1

…nd ECDSA-Full-R.

gstarovo force-pushed the ecdsa-sig-extend branch from 9f5252e to 4984db1 Compare March 26, 2026 10:49

new class RValue to propagate the ECPoint value to sigencode

76833c8

gstarovo force-pushed the ecdsa-sig-extend branch 2 times, most recently from 2949133 to 0a9aebd Compare March 26, 2026 13:29

fix the int/long type in python2.7

75a6f2e

gstarovo force-pushed the ecdsa-sig-extend branch from 0a9aebd to 75a6f2e Compare March 26, 2026 13:56

Conversation

gstarovo commented Oct 21, 2025

Uh oh!

Uh oh!

gstarovo commented Oct 21, 2025

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants