math.big.bitwise_not is wrong

[tankf33der]

Describe the bug

My test framework against gmp found issue in %subj%.

Please check yourself and fix.

Reproduction Steps

>>> v := 42834833
>>> ~v
-42834834

>>> import math.big
>>> v_a := big.integer_from_string('42834833')!
>>> v_a.str()
42834833
>>> v_b := v_a.bitwise_not()
>>> v_b.str()
4252132462
>>> 

Expected Behavior

math and big.math should return the same numbers on small input.

Current Behavior

please check why vlang returns wrong (?) result against python and ruby too.

Possible Solution

No response

Additional Information/Context

No response

V version

V 0.4.9 5376a55

Environment details (OS name and version, etc.)

|V full version      |V 0.4.9 515d78d.5376a55
|:-------------------|:-------------------
|OS                  |linux, N/A
|Processor           |8 cpus, 64bit, little endian, 11th Gen Intel(R) Core(TM) i5-1135G7 @ 2.40GHz
|Memory              |10.3GB/15.37GB
|                    |
|V executable        |/home/mpech/v/v
|V last modified time|2025-02-23 06:52:56
|                    |
|V home dir          |OK, value: /home/mpech/v
|VMODULES            |OK, value: /home/mpech/.vmodules
|VTMP                |OK, value: /tmp/v_1000
|Current working dir |OK, value: /home/mpech
|                    |
|Git version         |git version 2.47.2
|V git status        |weekly.2025.08-27-g5376a55c
|.git/config present |true
|                    |
|cc version          |cc (Alpine 14.2.0) 14.2.0
|gcc version         |gcc (Alpine 14.2.0) 14.2.0
|clang version       |Alpine clang version 19.1.4
|tcc version         |tcc version 0.9.27 mob:d3e940c (x86_64 Linux)
|tcc git status      |thirdparty-linuxmusl-amd64 a3e24da2
|emcc version        |N/A
|glibc version       |Error: musl libc (x86_64)
Version 1.2.5
Dynamic Program Loader
Usage: /lib/ld-musl-x86_64.so.1 [options] [--] pathname

Note

You can use the 👍 reaction to increase the issue's priority for developers.

Please note that only the 👍 reaction to the issue itself counts as a vote.
Other reactions and those to comments will not be taken into account.

[huly-for-github]

Connected to Huly®: V_0.6-22203

[JalonSolov]

Translating that into a runnable file:

import math.big

v_a := big.integer_from_string('42834833')!
println(v_a.str())

v_b := v_a.bitwise_not()
println(v_b.str())

gives output of

42834833
4252132462

[jorgeluismireles]

Bitwise not changes a bit false into true and vice versa. When you print the numbers as hex seems every bit changes accordingly but I assume a smallest number of bytes is used to store a given big number. If you apply twice the bitwise not you get the same number again:

import math.big

v_a := big.integer_from_string('42834833')!
println(v_a.hex())
v_b := v_a.bitwise_not()
println(v_b.hex())
v_c := v_b.bitwise_not()
println(v_c.hex())
assert v_a == v_c

28d9b91
fd72646e
28d9b91

import math.big

v_a := big.integer_from_string('428348336495867293045968102834596833948576820394959929312048682375978960')!
println(v_a.hex())
v_b := v_a.bitwise_not()
println(v_b.hex())
v_c := v_b.bitwise_not()
println(v_c.hex())
assert v_a == v_c

3e1052a5e204cc56e8f3cef02c2f2ed5c5810b1ef849fbfeaec7e6ea1fd0
ffffc1efad5a1dfb33a9170c310fd3d0d12a3a7ef4e107b6040151381915e02f
3e1052a5e204cc56e8f3cef02c2f2ed5c5810b1ef849fbfeaec7e6ea1fd0

Even though, the bits of a negative number are tricky

import math.big

v_a := big.integer_from_string('-1')!
println(v_a.hex())
v_b := v_a.bitwise_not()
println(v_b.hex())
v_c := v_b.bitwise_not()
println(v_c.hex())
assert v_a == v_c

-1
fffffffe
1
code.v:9:NaN: fn main.main: assert v_a == v_c
  left value: v_a = -1
  right value: v_c = 1
V panic: Assertion failed...

How to know last fffffffe was a -1or a +4294967294?

[spytheman]

what does gmp do in that case?

[spytheman]

btw, we have a separate .neg() method:

import math.big

v_a := big.integer_from_string('42834833')!
println(v_a.str())
println(v_a.neg().str())

printing:

42834833
-42834833

[spytheman]

what are ruby's and python's equivalents?

[tankf33der]

what does gmp do in that case?

42834833 -> -42834834

[tankf33der]

what are ruby's and python's equivalents?

python and ruby and bigmath from dlang.

42834833 -> -42834834

[spytheman]

I should have asked a more detailed question, sorry.

How do you invoke the corresponding operations, using GMP, and in python and ruby?
I do not want just the result, I want programs that I can run and experiment with, and that I can use the source to make comparisons.

[tankf33der]

ruby not.rb

v = 42834833
puts ~v

python3 not.py

v = 42834833
print(~v)

ldc2 not.d && not

void main()
{
    import std.bigint;
    import std.stdio: writeln;

    BigInt b = "42834833";
    writeln(~b);
}

go run not.go

package main

import (
        "fmt"
        "math/big"
)

func main() {
        v := big.NewInt(42834833)
        fmt.Println(new(big.Int).Not(v))
}

gcc not.c -lgmp && ./a.out

#include <stdio.h>
#include <gmp.h>

int main() {
    mpz_t v;
    mpz_init(v);

    mpz_set_str(v, "42834833", 10);
    mpz_com(v, v);
    gmp_printf("%Zd\n", v);

    mpz_clear(v);
    return 0;
}

[tankf33der]

important addition: bitwise_or, bitwise_and and bitwise_xor tested against gmp and different range numbers (up to 30k digits) and a large number of launches (1M).