Skip to content

Conversation

@miguel-ambrona
Copy link

PR on PSE:
privacy-ethereum#317

We disable two checks on key generation that made sure the row being filled was within the usable rows.

These checks are not necessary for fixed columns (where we disabled them), because fixed columns are public and are not blinded. Indeed, they are computed at setup time.

These checks served as a sanity check though, because in general one would not want to assign values at unusable offsets. On exception is the case of lookups. Thanks to this change, we can now run circuits with 2^k lookup tables on a k-domain. Before this change, a (k+1)-domain was needed.

This solution is good IMHO, but I suggest opening a GitHub issue to revisit this problem in the future to possibly enable/disable the check via an additional argument to the assign_fixed function.

@iquerejeta iquerejeta self-requested a review April 24, 2024 10:02
Copy link

@iquerejeta iquerejeta left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sounds good - could you open an issue to track the PR in PSE and to revisit this in the future?

@miguel-ambrona
Copy link
Author

miguel-ambrona commented Apr 24, 2024

Sounds good - could you open an issue to track the PR in PSE and to revisit this in the future?

Thanks @iquerejeta !
I'll open an issue to track that as soon as the issues are enabled for this repo.

@miguel-ambrona miguel-ambrona merged commit b410501 into fixed_73408a1 Apr 24, 2024
iquerejeta added a commit that referenced this pull request May 8, 2024
iquerejeta referenced this pull request in midnightntwrk/halo2 Apr 8, 2025
We do not add the architecture, as the verifier is already parametrised by that.
iquerejeta referenced this pull request in midnightntwrk/halo2 Apr 14, 2025
We do not add the architecture, as the verifier is already parametrised by that.
dkaidalov pushed a commit that referenced this pull request Jul 1, 2025
* Compare polynomials and polynomial commitments by reference.
Issue #1

Not permit duplicate queries
Issue #3

* Add Length of PIs to the transcript
Issue #2

* Check trailing bytes
Issue #6

* Panic if two points are equals in lagrange_interpolate.
Issue #10

* Hash the byte representation of the VK in the transcript Issue #5

We do not add the architecture, as the verifier is already parametrised by that.

* Consistency in polynomial multiplication Issue #7

We mutate the polynomial (with all zeroes) to be consistent with the behaviour.

* Make commitment homomorphic, and don't send evals of pieces - Issue#0
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants