Merge pull request #657 from kdmukai/decodeqr_bytes_fix_v2

[bugfix] `DecodeQR`: handle Compact SeedQR byte data that can be decoded to `str`

Author: newtonick

src/seedsigner/models/decode_qr.py

@@ -414,6 +414,14 @@ def detect_segment_type(s, wordlist_language_code=None):
             pass
 
         # Is it byte data?
+        if not isinstance(s, bytes):
+            try:
+                # TODO: remove this check & conversion once above cast to str is removed
+                s = s.encode()
+            except UnicodeError:
+                # Couldn't convert back to bytes; shouldn't happen
+                raise Exception("Conversion to bytes failed")
+
         # 32 bytes for 24-word CompactSeedQR; 16 bytes for 12-word CompactSeedQR
         if len(s) == 32 or len(s) == 16:
             try:

tests/test_seedqr.py

@@ -4,7 +4,6 @@
 from seedsigner.models.decode_qr import DecodeQR, DecodeQRStatus
 from seedsigner.models.encode_qr import SeedQrEncoder, CompactSeedQrEncoder
 from seedsigner.models.qr_type import QRType
-from seedsigner.models.seed import Seed
 
 
 
@@ -104,3 +103,28 @@ def test_compact_seedqr_handles_null_bytes():
     # 12-word seed, multiple null bytes in a row
     entropy = os.urandom(10) + b'\x00\x00' + os.urandom(4)
     run_encode_decode_test(entropy, mnemonic_length=12, qr_type=QRType.SEED__COMPACTSEEDQR)
+
+
+def test_compact_seedqr_bytes_interpretable_as_str():
+    """ 
+    Should successfully decode a Compact SeedQR whose bytes can be interpreted as a valid
+    string. Most Compact SeedQR byte data will raise a UnicodeDecodeError when attempting to
+    interpret it as a string, but edge cases are possible.
+
+    see: Issue #656
+    """
+    # Randomly generated to pass the str.decode() step; 12- and 24-word entropy.
+    entropy_bytes_tests = [
+        b'\x00' * 16,  # abandon * 11 + about
+        b'\x12\x15\\1j`3\x0bkL}f\x00ZYK',
+        b'tv\x1bZjmqN@t\x13\x1aK\\v)',
+        b'|9\x05\x1aHF9j\xda\xb6v\x05\x08#\x12=',
+        b"iHK`4\x1a5\xd3\xaf\xd3\xb47htJ.}<\xea\xbf\x88Xh\x01.?R2^\xc2\xb1'",
+        b'|Z\x11\x1dt\xdd\x97~t&f &G$H|^[\xd3\x9d<q]z\x14.\x11`!\xd1\x91',
+        b'0\xd4\xb3\\,\xcd\x8d7c/Rp\x0e\xc2\xbb\xe4\x99\xa3=j5,\xcc\x9a[>\x19Z{\ng^',
+    ]
+
+    for entropy_bytes in entropy_bytes_tests:
+        entropy_bytes.decode()  # should not raise an exception
+        mnemonic_length = 12 if len(entropy_bytes) == 16 else 24
+        run_encode_decode_test(entropy_bytes, mnemonic_length=mnemonic_length, qr_type=QRType.SEED__COMPACTSEEDQR)