Fix modpow for exp=1 #113
Fix modpow for exp=1 #113
Conversation
There was a problem hiding this comment.
OK, so this is a regression in 0.2.3 following the optimizations in #93. While I expect a 1 exponent would be unusual, we should still fix this, and I appreciate that you're proactive here. Let's get this sorted out and published!
src/biguint.rs
Outdated
| @@ -2285,6 +2285,9 @@ fn plain_modpow(base: &BigUint, exp_data: &[BigDigit], modulus: &BigUint) -> Big | |||
|
|
|||
| let mut exp_iter = exp_data[i + 1..].iter(); | |||
| if exp_iter.len() == 0 && r.is_one() { | |||
| if &base > modulus { | |||
| base %= modulus; | |||
| } | |||
There was a problem hiding this comment.
I think it would be better to normalize this above where we create a base.clone(), just use base % modulus instead. The implementation of that % operator already checks if the value is less than the modulus too.
| @@ -81,6 +81,7 @@ mod biguint { | |||
| check_modpow::<u32>(0, 15, 11, 0); | |||
| check_modpow::<u32>(3, 7, 11, 9); | |||
| check_modpow::<u32>(5, 117, 19, 1); | |||
| check_modpow::<u32>(20, 1, 2, 0); | |||
There was a problem hiding this comment.
Please add a similar test with an odd modulus too, so that corner case is not missed. (But I think that implementation is already OK.)
There was a problem hiding this comment.
Additionally, I added api surface test for small values below
youknowone
force-pushed
the
modpow-exp1
branch
2 times, most recently
from
73731c8
to
904f057
Compare
youknowone
force-pushed
the
modpow-exp1
branch
from
904f057
to
00042ec
Compare
|
Thanks! bors r+ |
Build succeeded |
Unfortunately, there was cases modulus never happened