How does bitwise 'or' becomes a '+=' in bytecode for python

Question

I was reading the following explanation of python bytecode. I'm posting a lot in order to give the full context. I'm confused about the part where the bitwise 'or' becomes a '+='.

EXTENDED_ARG 1
CALL_FUNCTION 4
When the interpreter executes EXTENDED_ARG, its oparg (which is 1) is left-shifted by eight bits and stored in a temporary variable. Let’s call it extended_arg (do not confuse it with the opname EXTENDED_ARG):
extened_arg = 1 << 8  # same as 1 * 256
So the binary value 0b1 (the binary value of 1) is converted to 0b100000000. This is like multiplying 1 by 256 in the decimal system and extened_arg will be equal to 256. Now we have two bytes in extened_arg. When the interpreter reaches to the next instruction, this two-byte value is added to its oparg (which is 4 here) using a bitwise or.
extened_arg = extened_arg | 4
# Same as extened_arg += 4
This is like adding the value of the oparg to extened_arg. So now we have:
extened_arg = 256 + 4 = 260
and this value will be used as the actual oparg of CALL_FUNCTION. So, in fact,
EXTENDED_ARG 1
CALL_FUNCTION 4
is interpreted as:
EXTENDED_ARG 1
CALL_FUNCTION 260

Can someone explain how this translates.

The bitwise-or of `4` and `256` happens to be equal to the sum of `4` and `256`, because they have no set bits in common. This has nothing to do with either Python or bytecode; it's a pure *mathematical* fact. — Karl Knechtel, Oct 16 '22 at 03:28

score 2 · Answer 1 · answered Dec 05 '20 at 11:57

2

If x and y share no set bits, then x | y is equivalent to x + y.

answered Dec 05 '20 at 11:57

user2357112

260,549
28
431
505

What does it mean to share a set bit? – Joemoor94 Dec 05 '20 at 11:58
@Joemoor94 `x << 8` (where *x* is 1 but whatever) and 4 do not have common bits because 4 is less than 256 – bruno Dec 05 '20 at 12:01
@JoeyMoore: Line up the binary representations of the numbers (padded to the same length) and look for any columns where both have a 1. 1s in the binary representation are set bits. – user2357112 Dec 05 '20 at 12:01
@bruno I think this is highlighting a deeper ignorance on my part – Joemoor94 Dec 05 '20 at 12:05

Nine Tails · Answer 2 · 2020-12-05T12:18:58.647

1

step 1 : 1 << 8 this is equal to 100000000.
Which is 256 in decimal
step 2: now 256 + 4 = 260. (obviously)
256|4 is same as 100000000|100 = 100000100 = 260. (check how bitwise operators work)

from the above 2 statements the author concluded that += is same as |

note about or operator:
if you want to find the result of ORing to operands, just line them up, like you do when you add 2 numbers, then look at each pair of bits from right to left and if there is at least one 1 bit then the resulting number will have a 1 bit too, otherwise it is a 0.

edited Dec 05 '20 at 12:18

answered Dec 05 '20 at 12:09

Nine Tails

378
3
15

warning in the current conditions + is same as |, not += which is same as |= – bruno Dec 05 '20 at 12:11

score 0 · Answer 3 · answered Dec 05 '20 at 12:28

0

If one of the operands is known to be zero, then 0 + bit = bit gives the same result as 0 | bit == bit, where bit is 0 or 1.

answered Dec 05 '20 at 12:28

VPfB

14,927
6
41
75

How does bitwise 'or' becomes a '+=' in bytecode for python

3 Answers3