http://docs.python.org/3.3/library/stdtypes.html?highlight=to_bytes#int.to_bytes and http://docs.python.org/3.3/library/stdtypes.html?highlight=to_bytes#int.to_bytes would benefit from an example showing what they do based on simpler coding.

I have such an example that I wrote here: http://paddy3118.blogspot.co.uk/2012/11/some-identities-for-python-inttobytes.html that you can use.

I.e.

>> n = 2491969579123783355964723219455906992268673266682165637887
>> length = 25
>> n2bytesbig = n.to_bytes(length, 'big')
>> n2byteslittle = n.to_bytes(length, 'little')
>> assert n2bytesbig == bytes( (n >> i*8) & 0xff for i in reversed(range(length)))
>> assert n2byteslittle == bytes( (n >> i*8) & 0xff for i in range(length))
>> assert n == sum( n2bytesbig[::-1][i] << i*8 for i in range(length) )
>> assert n == sum( n2byteslittle[i] << i*8 for i in range(length) )
>> assert n == int.from_bytes(n2bytesbig, byteorder='big')
>> assert n == int.from_bytes(n2byteslittle, byteorder='little')
>>

Your example is comprehensive but not simple and obvious.
I think better to keep it out of doc.

I agree. The examples in the doc seem clear to me, whereas the ones you proposed are not as clear. Do you think there's something that they don't currently cover that should be added?

On 06/12/2012 14:31, Ezio Melotti wrote:

Ezio Melotti added the comment:

I agree. The examples in the doc seem clear to me, whereas the ones you proposed are not as clear. Do you think there's something that they don't currently cover that should be added?

----------
nosy: +ezio.melotti

---

Python tracker <report@bugs.python.org>
<http://bugs.python.org/issue16580\>

---

First, Thanks Ezio and Andrew for your replies.

My problem was that when working on bitcoin address validation I saw
code that was shifting and &'ing with 0xFF to convert to multiple bytes
and half remembered that there might be a Python function to do that. On
finding the .to_bytes method and its parameter "big" or "little", the
only way I had of working out which to use was to try each until I found
out which worked.

I therefore thought that what would have helped me was code that showed
the equivalent "expanded Python" for the method in a similar way to what
is done for some of the itertools functions etc.

If we split my request into two:

1. Is such extra explanation necessary.
2. Is my specific code that extra explanation.

I can work on the code a bit more.
Have I persuaded you that an extra explanation is necessary?

Thanks, Paddy.

P.S. I guess what is currently present shows the result of the methods
but nothing on how it could be generated. I am stating that the
generation can aid comprehension.

Usually we add plain Python equivalents when they are simple enough that the code equivalent is as understandable as the prose or more (see for example http://docs.python.org/3/library/functions.html#all, or the itertools functions you mentioned).
For this case I think it would help if you presented an equivalent function, e.g.:

def to_bytes(n, length, order):
    if order == 'little':
        return bytes((n >> i*8) & 0xff for i in range(length))
    elif order == 'big':
        return bytes((n >> i*8) & 0xff for i in reversed(range(length)))

or even:

def to_bytes(n, length, order):
    indexes = range(length) if order == 'little' else reversed(range(length))
    return bytes((n >> i*8) & 0xff for i in indexes)

This is also done for http://docs.python.org/3.3/library/stdtypes.html#int.bit_length just above to/from_bytes, so it might be a good addition.
If this is done, the equivalent function can also be added to the test suite, so we can verify that it's indeed equivalent.

On 09/12/2012 10:55, Ezio Melotti wrote:

Ezio Melotti added the comment:

Usually we add plain Python equivalents when they are simple enough that the code equivalent is as understandable as the prose or more (see for example http://docs.python.org/3/library/functions.html#all, or the itertools functions you mentioned).
For this case I think it would help if you presented an equivalent function, e.g.:

def to_bytes(n, length, order):
if order == 'little':
return bytes((n >> i*8) & 0xff for i in range(length))
elif order == 'big':
return bytes((n >> i*8) & 0xff for i in reversed(range(length)))

or even:

def to_bytes(n, length, order):
indexes = range(length) if order == 'little' else reversed(range(length))
return bytes((n >> i*8) & 0xff for i in indexes)

This is also done for http://docs.python.org/3.3/library/stdtypes.html#int.bit_length just above to/from_bytes, so it might be a good addition.
If this is done, the equivalent function can also be added to the test suite, so we can verify that it's indeed equivalent.

----------
keywords: +easy
stage: -> needs patch
versions: +Python 2.7, Python 3.2, Python 3.4

---

Python tracker <report@bugs.python.org>
<http://bugs.python.org/issue16580\>

---

The second example looks great. I like the dual use for testing too and
will try and remember both the next time I find I have ireas about the
documentation.

Thanks guys. It's appreciated!

Patch adding examples + tests for equivalence. Comments appreciated.

In particular, I'm not sure that the from_bytes example is simple enough to be useful:

def from_bytes(bytes, byteorder, signed=False):
    if byteorder == 'little':
        little_ordered = list(bytes)
    elif byteorder == 'big':
        little_ordered = list(reversed(bytes))
     n = sum(little_ordered[i] << i*8 for i in range(len(little_ordered)))
    if signed and little_ordered and (little_ordered[-1] & 0x80):
         n -= 1 << 8*len(little_ordered)
    return n

I don't call it the examples. Here is the examples:

>>> (1234).to_bytes(2, 'big')
b'\x04\xd2'
>>> (1234).to_bytes(2, 'little')
b'\xd2\x04'
>>> (-1234).to_bytes(2, 'big', signed=True)
b'\xfb.'
>>> int.from_bytes(b'\xde\xad\xbe\xef', 'big')
3735928559
>>> int.from_bytes(b'\xde\xad\xbe\xef', 'little')
4022250974
>>> int.from_bytes(b'\xde\xad\xbe\xef', 'big', signed=True)
-559038737

The from_bytes example code looks fine to me, except that I'd probably use enumerate in the iteration; i.e.,

    sum(b << 8*i for i, b in enumerate(little_ordered))

instead of

    sum(little_ordered[i] << i*8 for i in range(len(little_ordered)))

Also, in:

   if signed and little_ordered and (little_ordered[-1] & 0x80):

I wondered why you needed the little_ordered check. But I see that int.from_bytes(b'', 'little', signed=True) produces 0, which is a little bit disappointing: I was expecting an exception. (A signed format should have a sign bit, which is impossible if the length of the byte string is 0.)

The to_bytes example code is missing range checks for the input. It may be clearer to simply state that in the docs, instead of modifying the example code to add the range checks.

I've rewritten woparry's patch so that it targets the current codebase as well as fixing the issues that Mark mentioned. It appears that the range checks on to_bytes have already been added to the documentation so I haven't added those.

Should a separate issue be raised for the behaviour of int.from_bytes(b'', <byteorder>, signed=True')? I think it's out of scope for this issue.

@Gautam

I think it's out of scope for this issue.

Agreed. It was just a note in passing. I don't think it's worth opening a new issue, though, unless this is genuinely causing problems for someone. Just let it lie.

New changeset ad0a8a9 by Gautam Chaudhuri in branch 'main':
bpo-16580: [doc] Add examples to int.to_bytes and int.from_bytes (GH-27760)
ad0a8a9

Merged. Thank you for the contribution!