Error in the specification of the sparse_factorize function.
Closed this issue · 12 comments
The documentation says that w is the first column of m_hat or second column of m without the first row..
Actually it is the first column of m without the first row.
The implementation is ok: https://github.com/filecoin-project/neptune/blob/bafd77a5014e3b6a6b40359097835c3eb1dd533f/src/mds.rs#L197
I think this refers to the spec (https://github.com/filecoin-project/specs). @jbaylina Can you file there and tag @DrPeterVanNostrand who can fix? We can close this once it's accounted for there — unless I've misunderstood the issue.
I'll post there. But in this repo there is a pdf with spec. https://github.com/filecoin-project/neptune/blob/master/poseidon_spec.pdf
Ah, right. I believe that is generated from the spec. So once the spec is updated, we should regenerate the PDF and update here as well. Thanks.
The change has also been added to the Filecoin spec here: filecoin-project/specs#1278
@jbaylina I didn't see your specs PR before submitting my own
The PDF in this repo is actually rendered using an app called Typora
The PDF in this repo is actually rendered using an app called Typora
I see. If the source isn't directly extracted from the spec, we should probably commit it along with (at least some manual) build instructions.
@porcuquine I've had the markdown in a local git repo since it was published. I just made a public git repo here: https://github.com/DrPeterVanNostrand/poseidon-spec
Basically, I make the necessary changes to my local markdown version, then render as a PDF via Typora, then PR the PDF to neptune. Then I PR the changes in MathJax to the Filecoin specs repo.
Got it. Do you mind just writing those instructions into the source repo (or make a directory for it here). That way if someone else needs to do it at some point, we at least know both how to produce the artifact and to make sure everything is in sync.
Ok, I added the Markdown file and rendering instructions to PR #114
Merged.