格式化python unittests的测试失败

Question

I am running a number of unittests to test a binary protocol:

An example failure messages is as follows:

AssertionError: Items in the first set but not the second:
b'\x00\x02@=\x00'
Items in the second set but not the first:
b'\x00\x02@N\x00'

This is unwieldy as I need to convert the ascii chars into hex manually to check what is happening.

It would be nice, if it was possible to cause unittest to format all bytes objects as hex, eg

AssertionError: Items in the first set but not the second:
b'\x00\x02\x40\x3d\x00'
Items in the second set but not the first:
b'\x00\x02\x40\x4e\x00'

Any advice on how to accomplish this with the lease amount of effort?

Note: I have not only this particular comparison between two set, but also comparisons between lists and dicts... Hence my request for a low-effort solution.

Answer 1

Ok, I dug into the code of unittest myself.

Internally, it uses repr(obj) to format the output of test failures. Short of rewriting all the comparison routines for list, set, dict, etc. there is no easy fix.

The best solution I found is subclassing the bytes object and overloading its __repr__ method like so:

class st_frame(bytes):
    def __repr__(self):
        return "b'{}'".format(
            ''.join(['\\x%02x' % c for c in self])
        )

Then one can wrap the results with this object within the test cases:

ret = func_to_test(data)      # lets assume this returns list of bytes objects
ret = [st_frame(value) for value in ret]
self.assertEqual(ret, target)