Const conversions Uint128 -> Uint256 and Uint256 -> Uint512 #1110
Conversation
|
Looking commit by commit might make this a bit clearer in this case, I think. ...except for the part where I forgot to update the commit msg. Ahem. |
packages/std/src/math/uint256.rs
Outdated
| let words: [u64; 4] = [ | ||
| u64::from_le_bytes([ | ||
| bytes[0], bytes[1], bytes[2], bytes[3], bytes[4], bytes[5], bytes[6], bytes[7], | ||
| ]), | ||
| u64::from_le_bytes([ | ||
| bytes[8], bytes[9], bytes[10], bytes[11], bytes[12], bytes[13], bytes[14], | ||
| bytes[15], | ||
| ]), | ||
| 0, | ||
| 0, | ||
| ]; | ||
| Self(U256(words)) |
There was a problem hiding this comment.
Wouldn't it be nice to reuse Self::from_be_bytes here? This reduces the places where we mess with U256 internals.
packages/std/src/math/uint512.rs
Outdated
| 0, | ||
| ]; | ||
|
|
||
| Self(U512(words)) |
There was a problem hiding this comment.
And with #1111 we can do the same here too, right?
There was a problem hiding this comment.
I wonder if that would introduce some overhead though when this function isn't used in a const context. Does it matter? Probably not.
There was a problem hiding this comment.
Good question. Compared to every serialization/deserialization all this conversion is trivial. I think it would be good to avoid exposing implementation details here. Ever Uint* type should have efficient BE/LE exporters and importers. Let's focus on those.
| bytes[8], bytes[9], bytes[10], bytes[11], bytes[12], bytes[13], bytes[14], bytes[15], | ||
| 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, | ||
| ]) | ||
| } |
uint
force-pushed
the
const-conversions
branch
from
dce543f to
fe9d419
Compare
|
Rebased. |
|
Do you implement |
That's exactly what I was doing 👍 |
Closes #1107