The test suite only uses hardcoded test fixtures. There are no round-trip tests with random data (including leading zero bytes) that would catch encoding/decoding bugs across all 24 encodings. go-multibase has a comprehensive TestRoundTrip that does this.
Problem
Current tests use 85 hardcoded (encoding, data, expected) tuples. These cover common cases but miss:
- Leading zero bytes —
\x00\x00hello (catches Issue 1)
- All-zeros data —
\x00\x00\x00\x00
- All-ones data —
\xff\xff\xff\xff
- Random data of varying lengths — 1 byte, 16 bytes, 137 bytes
- Large data — 1024+ bytes
go-multibase generates random data and tests every encoding:
func TestRoundTrip(t *testing.T) {
buf := make([]byte, 137+16)
rand.Read(buf[16:]) // 16 leading zeros + 137 random bytes
for base := range EncodingToStr {
for i := 0; i < len(buf); i++ {
// Test from full buffer down to single byte
enc, _ := Encode(base, buf[i:])
_, out, _ := Decode(enc)
if !bytes.Equal(buf[i:], out) {
t.Fatal("round-trip failed")
}
}
}
}Proposed Solution
Add tests/test_roundtrip.py:
import os
import pytest
from multibase import encode, decode, ENCODINGS
@pytest.mark.parametrize("encoding_info", ENCODINGS, ids=lambda e: e.encoding)
class TestRoundTrip:
def test_random_data(self, encoding_info):
"""Round-trip random data of various sizes."""
for size in [1, 2, 7, 16, 32, 64, 137, 256, 1024]:
data = os.urandom(size)
encoded = encode(encoding_info.encoding, data)
decoded = decode(encoded)
assert decoded == data, f"Failed for {encoding_info.encoding} size={size}"
def test_leading_zeros(self, encoding_info):
"""Round-trip data with leading zero bytes."""
for num_zeros in [1, 2, 4, 8, 16]:
data = b'\x00' * num_zeros + b'hello'
encoded = encode(encoding_info.encoding, data)
decoded = decode(encoded)
assert decoded == data, \
f"Leading zeros lost for {encoding_info.encoding} zeros={num_zeros}"
def test_all_zeros(self, encoding_info):
"""Round-trip all-zero data."""
for size in [1, 4, 16, 32]:
data = b'\x00' * size
encoded = encode(encoding_info.encoding, data)
decoded = decode(encoded)
assert decoded == data
def test_all_ones(self, encoding_info):
"""Round-trip all-0xFF data."""
for size in [1, 4, 16, 32]:
data = b'\xff' * size
encoded = encode(encoding_info.encoding, data)
decoded = decode(encoded)
assert decoded == dataRelated
The test suite only uses hardcoded test fixtures. There are no round-trip tests with random data (including leading zero bytes) that would catch encoding/decoding bugs across all 24 encodings. go-multibase has a comprehensive
TestRoundTripthat does this.Problem
Current tests use 85 hardcoded
(encoding, data, expected)tuples. These cover common cases but miss:\x00\x00hello(catches Issue 1)\x00\x00\x00\x00\xff\xff\xff\xffgo-multibase generates random data and tests every encoding:
Proposed Solution
Add
tests/test_roundtrip.py:Related
multibase_test.goTestRoundTrip