MASP CI False Positives and Negatives

[murisi]

The CI reports that the 3 MASP E2E tests are passing when they should be failing. That the tests should be failing can be established from the fact that trying to manually construct shielded-to-shielded transactions with namadac (almost?) always fails due to issues with the MASP transaction signer, and also because the MASP E2E tests (almost?) always fail when executed locally.

Interestingly when the aforementioned issues with namadac are fixed, the local E2E tests start passing but the 3 MASP E2E tests in the CI begin to fail. For reference, these fixes are at

namada/apps/src/lib/client/signing.rs

Line 157 in 5504002

TxSigningKey::SecretKey(signing_key) => signing_key,

and

namada/apps/src/lib/client/tx.rs

Line 1909 in 5504002

default_signer.clone(),

.

A solution to this issue would prevent future regressions in the MASP code. cc @Fraccaman @gijswijs

[cwgoes]

Closing since this seems to be out-of-date, please reopen if I'm mistaken.