Add Encryption and Decryption for OPRF Report Structs #904
Conversation
|
I will use slices instead of generic arrays to fix the trait bounds. So don't review it yet. |
I fixed it, its ready for review now. |
|
Oh, and I'd suggest either cleaning up the commit history locally (I use |
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #904 +/- ##
=======================================
Coverage ? 90.29%
=======================================
Files ? 182
Lines ? 26091
Branches ? 0
=======================================
Hits ? 23560
Misses ? 2531
Partials ? 0 ☔ View full report in Codecov by Sentry. |
|
This PR hasn't seen any updated in 2 weeks, are we still planning to work on it? |
ipa-core/src/report.rs
Outdated
| ) | ||
| .unwrap(); // validated on construction | ||
|
|
||
| let mut ct_mk = self.mk_ciphertext().to_vec(); |
There was a problem hiding this comment.
Can you point me to the heap allocation in the old code? I am struggling to find it - it does seem that it works w/o a roundtrip to jemalloc.
The reason I think it matters is because this will be called 50 million (or whatever shard size is) times - we got a decent chance for memory allocator to do its job properly, but this is a game I think we don't have to play as it should be possible to avoid in this code
|
|
||
| let mut ct_mk: GenericArray<u8, CTMKLength> = | ||
| *GenericArray::from_slice(self.mk_ciphertext()); | ||
| // let mut ct_mk = self.mk_ciphertext().to_vec(); |
There was a problem hiding this comment.
don't want to hold this change for too long, but lets make sure we clean up commented code. Fine with a follow up PR to address those
| let mut ct_btt: GenericArray<u8, CTBTTLength<BK, TV, TS>> = | ||
| GenericArray::from_slice(self.btt_ciphertext()).clone(); |
There was a problem hiding this comment.
you can use the same dereferencing trick *GenericArray instead of GenericArray(...).clone()
No description provided.