Consider moving return value witnesses directly after the inputs so function signature fully defines ABI #5104
Comments
|
This will make it so param_witnesses and return_witnesses are not needed anymore right? since they can be completely derived from the types |
|
Yep, that's correct. |
5 tasks
github-merge-queue bot
pushed a commit
that referenced
this issue
May 31, 2024
…5142) # Description ## Problem\* Resolves #5104 ## Summary\* This PR preallocates some witnesses to hold the return values at the beginning of ACIR gen and then adds assertions to fill these witnesses with the return values. This ensures that the return values will be placed in the witness map directly after any function inputs (reasons for this being desirable are laid out in #5104) ## Additional Context ## Documentation\* Check one: - [x] No documentation needed. - [ ] Documentation included in this PR. - [ ] **[For Experimental Features]** Documentation to be submitted in a separate PR. # PR Checklist\* - [x] I have tested the changes locally. - [x] I have formatted the changes with [Prettier](https://prettier.io/) and/or `cargo fmt` on default settings. --------- Co-authored-by: jfecher <jake@aztecprotocol.com>
5 tasks
github-merge-queue bot
pushed a commit
that referenced
this issue
Jun 10, 2024
# Description ## Problem\* Followup to #5142 Related to #5104 ## Summary\* This PR deletes the `param_witnesses` and `return_witnesses` fields from the ABI as the locations of all inputs/return values are defined implicitly by the types themselves. I've left this as draft as there's some changes which need to be made to the JS code which will require some extra work (it needs to be made aware of the sizes of various types.) ## Additional Context ## Documentation\* Check one: - [ ] No documentation needed. - [x] Documentation included in this PR. - [ ] **[For Experimental Features]** Documentation to be submitted in a separate PR. # PR Checklist\* - [x] I have tested the changes locally. - [x] I have formatted the changes with [Prettier](https://prettier.io/) and/or `cargo fmt` on default settings.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
We recently moved to a situation where we always generate distinct witnesses for all of the return values (#4914). This means that we can know exactly how many witnesses will be needed for the return values before performing acir gen.
This then raises the potential for just reserving these witness indices so that instea of
[...program inputs, ...intermediate witnesses, ...program return values]the witnesses are laid out as[...program inputs, ...program return values, ...intermediate witnesses].The result of this is that ABI encoding/decoding of arguments/return values will be entirely defined by the function signature and independent of the rest of the circuit. This will allow us to have a
foundry castlike tool for ABI encoding/decoding witnesses.The text was updated successfully, but these errors were encountered: