Clean up reject_callx_r10

[ripatel-fd]

Problem

Firedancer always cleans up features after Agave does.
We do so to ensure latest version Firedancer can replay a ledger produced by latest version Agave.

We'd like to see reject_callx_r10 (activated on all networks) cleaned up in Agave to allow Firedancer to also clean up the feature.

Summary of Changes

Hardcodes reject_callx_r10 to on.

Fixes #

[ripatel-fd]

@Lichtso There is a loaded program cache test built around the reject_callx_r10 feature. How do you suggest we fix the test I broke?

[Lichtso]

Ah yes, we need a different test program and a different feature. Currently it uses reject_callx_r10 because it is easy to have a program which succeeds or fails the verifier depending on the status of the feature.

[mergify]

The Firedancer team maintains a line-for-line reimplementation of the
native programs, and until native programs are moved to BPF, those
implementations must exactly match their Agave counterparts.
If this PR represents a change to a native program implementation (not
tests), please include a reviewer from the Firedancer team. And please
keep refactors to a minimum.

[Lichtso]

@ripatel-fd In test_feature_activation_loaded_programs_epoch_transition and test_deploy_last_epoch_slot can you replace reject_callx_r10 with disable_fees_sysvar? I think that should suffice.

[ripatel-fd]

@Lichtso Done, getting this test failure:

---- bank::tests::test_feature_activation_loaded_programs_cache_preparation_phase stdout ----
thread 'bank::tests::test_feature_activation_loaded_programs_cache_preparation_phase' panicked at runtime/src/bank/tests.rs:11947:5:
assertion `left == right` failed
  left: Err(InstructionError(0, InvalidAccountData))
 right: Err(InstructionError(0, ProgramFailedToComplete))

[ripatel-fd]

Well, makes sense. The test binary is called callx-r10-sbfv1.so. I can write a new test that attempts to read the fees sysvar.

[Lichtso]

I can write a new test that attempts to read the fees sysvar.

Not necessary anymore, I found a cleaner solution. Instead of continuing to kick the can down the street and updating that program manually each time we clean a verifier-affecting feature we introduce a generally program rejecting test feature.

[ripatel-fd]

Not necessary anymore, I found a cleaner solution. Instead of continuing to kick the can down the street and updating that program manually each time we clean a verifier-affecting feature we introduce a generally program rejecting test feature.

That sounds great, thanks. I'll mark this PR as a draft then and wait for that test feature to be merged. Or maybe you can just copy over my patch to your branch to save time...

[Lichtso]

Resolved in #2884