Distinguish between integer, field element and bool. #910
Conversation
chriseth
force-pushed
the
integers
branch
7 times, most recently
from
765b9b8 to
f96cdfc
Compare
chriseth
force-pushed
the
integers
branch
3 times, most recently
from
7b1bfea to
40ba2f0
Compare
chriseth
force-pushed
the
integers
branch
2 times, most recently
from
8c3274e to
47ae649
Compare
chriseth
force-pushed
the
builtins
branch
2 times, most recently
from
c5e5597 to
3e3349d
Compare
chriseth
changed the title
chriseth
force-pushed
the
integers
branch
2 times, most recently
from
c3d49ed to
4811624
Compare
| @@ -229,23 +232,23 @@ namespace Arith(N); | |||
| // I have no idea where the "- 4 * prepend_zeros(p, 16)" comes from, but it was in the original constraints. | |||
| let product_with_p = (|| |x| |nr| product(p, x)(nr) - 4 * prepend_zeros(p, 16)(nr))(); | |||
|
|
|||
| let eq1 = (|| |nr| product(sf, x2f)(nr) - product(sf, x1f)(nr) - y2f(nr) + y1f(nr) + product_with_p(q0f)(nr))(); | |||
| let eq1 = (|| |nr| std::field::modulus() + product(sf, x2f)(nr) - product(sf, x1f)(nr) - y2f(nr) + y1f(nr) + product_with_p(q0f)(nr))(); | |||
There was a problem hiding this comment.
Actually I need to check if we still need this.
book/src/pil/types.md
Outdated
|
|
||
| Powdr-pil is usually side-effect free, but there are some situations where side-effects happen. | ||
| This includes the `std::debug::print` built-in function. Expressions are usually evaluated | ||
| left-to right and eagerly, but it is best to not rely on this behaviour. |
There was a problem hiding this comment.
I think we should be exhaustive on the exceptions
book/src/pil/types.md
Outdated
| The lambda expression `|| 7` is a function that returns a constant (has no parameters). | ||
| Lambda functions can capture anything in their environment and thus form closures. | ||
|
|
||
| Due to the dynamic typing of powdr-pil, the exact type of a lambda expression can often not be determined. |
There was a problem hiding this comment.
This sounds confusing. Why does it do dynamic typing?
There was a problem hiding this comment.
Because nobody found the time to implement a proper generic type checker. This will be replaced once we have a proper type checker. Thibaut told me that I should not document what we plan but what we have...
There was a problem hiding this comment.
yea but still, we don't do dynamic typing, it's all compile-time
There was a problem hiding this comment.
Ok, it depends on what you mean by compiling. But even witness generation uses dynamic typing since you can call functions from the query callbacks and those functions could look like
|i| match i {
0 => [],
1 => 7,
_ => (1,2)
}
There was a problem hiding this comment.
wat, really?! Can we stop that :D
chriseth
force-pushed
the
integers
branch
8 times, most recently
from
1713e69 to
007b605
Compare
test_data/std/poseidon_gl_test.asm
Outdated
| @@ -55,7 +55,7 @@ machine Main { | |||
| assert_eq C, 14940024623104903507; | |||
| assert_eq D, 10772674582659927682; | |||
|
|
|||
| A, B, C, D <== poseidon(18446744069414584321, 18446744069414584321, 18446744069414584321, 18446744069414584321, 18446744069414584321, 18446744069414584321, 18446744069414584321, 18446744069414584321, 0, 0, 0, 0); | |||
| A, B, C, D <== poseidon(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0); | |||
There was a problem hiding this comment.
Yes, the number is the goldilocks modulus. So the test should do exactly the same as it did before.
@georgwiese any reason you chose these numbers (and used explicit zeros for the last four elements)?
There was a problem hiding this comment.
These test vectors are the same as polygon's. I guess they were chosen exactly to test that the result is the same as if you change the inputs to 0s. The last 4 are 0s because they're the cap that are used to "change" the hash function
There was a problem hiding this comment.
Exactly, they come from the JSON linked in the comment a few lines above.
chriseth
force-pushed
the
integers
branch
4 times, most recently
from
a6075b9 to
5532b6a
Compare
No description provided.