refactor: tagging related cleanup #17451

benesjan · 2025-10-02T09:30:53Z

I started working on improving unconstrained tagging as proposed in this forum post and the initial step in implementing that is the following:

store the tagging indexes requested during execution in a cache (just like we do with notes),
when the execution is finished take the cache and along with a tx hash dump it into db.

Given that the cache will reside in PrivateExecutionOracle I now need to orchestrate the sync there.

When trying to do that (to be done in a PR up the stack) I realized that the current tagging-related API was cumbersome and this made me do the following in this PR:

Have the API accept directionalAppTaggingSecret on the input instead of appTaggingSecret and recipient as now I can just compute it in the PrivateExecutionOracle and request info based on that,
I moved the tagging functionality from stdlib to PXE as there was no reason to have it in stdlib,
I simplified a bunch of random things.

This PR is a bit random so if something is not clear LMK.

benesjan · 2025-10-02T09:31:10Z

yarn-project/pxe/src/contract_function_simulator/oracle/private_execution_oracle.ts

yarn-project/stdlib/src/logs/indexed_tagging_secret.ts

yarn-project/stdlib/src/keys/derivation.ts

yarn-project/pxe/src/contract_function_simulator/execution_data_provider.ts

yarn-project/pxe/src/contract_function_simulator/pxe_oracle_interface.ts

yarn-project/pxe/src/contract_function_simulator/tagging_utils.ts

yarn-project/pxe/src/storage/tagging_data_provider/tagging_data_provider.ts

yarn-project/pxe/src/contract_function_simulator/tagging_utils.ts

yarn-project/pxe/src/contract_function_simulator/execution_data_provider.ts

benesjan · 2025-10-06T10:18:04Z

yarn-project/stdlib/src/logs/directional_app_tagging_secret.ts

+   * @param localAddress - The complete address of entity 1 in the shared tagging secret derivation scheme
+   * @param localIvsk - The incoming viewing secret key of entity 1


Compared to the original code I prefixed these 2 with "local" to communicate that they correspond to 1 entity. Then I think it's clearer that external corresponds to the entity 2.

benesjan · 2025-10-06T10:30:19Z

yarn-project/stdlib/src/logs/directional_app_tagging_secret.ts

+  externalAddress: AztecAddress,
+): Promise<Point> {
+  const knownPreaddress = await computePreaddress(await localAddress.publicKeys.hash(), localAddress.partialAddress);
+  // TODO: #8970 - Computation of address point from x coordinate might fail


Copied this TODO here with the original code. Josie tackled that issue in this PR and I think we can afford to remove it from here as it seems fine that we throw in the toAddressPoint function on the next line if the address is invalid here.

Agree?

benesjan · 2025-10-08T10:29:57Z

@nventuro managed to not rebase this with graphite so looking for changes since your last review should work.

I started working on improving unconstrained tagging as proposed in [this forum post](https://forum.aztec.network/t/on-note-discovery-and-index-coordination/7165) and the initial step in implementing that is the following: 1. store the tagging indexes requested during execution in a cache (just like we do with notes), 2. when the execution is finished take the cache and along with a tx hash dump it into db. Given that the cache will reside in `PrivateExecutionOracle` I now need to orchestrate the sync there. When trying to do that (to be done in a [PR up the stack](#17445)) I realized that the current tagging-related API was cumbersome and this made me do the following in this PR: 1. Have the API accept directionalAppTaggingSecret on the input instead of appTaggingSecret and recipient as now I can just compute it in the `PrivateExecutionOracle` and request info based on that, 2. I moved the tagging functionality from stdlib to PXE as there was no reason to have it in stdlib, 3. I simplified a bunch of random things. This PR is a bit random so if something is not clear LMK.

benesjan marked this pull request as draft October 2, 2025 10:20

benesjan force-pushed the 10-02-refactor_tagging_related_cleanup branch from c8c5c4f to bcabcea Compare October 2, 2025 10:21