v3.1: runtime: resolve compute budget in check_transaction() (backport of #9155)

[mergify]

Problem

in implementing simd290 and simd297 it would be convenient if the only errors we had to handle in svm validate_transaction_nonce_and_fee_payer() were the nonce and feepayer errors because we will have to propagate particular representations for those cases upwards

the only error other than nonce/feepayer is unwrapping Result<SVMTransactionExecutionAndFeeBudgetLimits>, constructed in runtime check_transactions()

Summary of Changes

instead of delivering a malformed compute budget as its own error, transform it into a Result<CheckedTransactionDetails> error in runtime. the only possible errors are in sanitize_and_convert_to_compute_budget_limits(), checking that requested data and heap size values are sane

there are tests in several crates for a DuplicateInstruction error, which is not relevant to this pr. this type of error gets resolved much earlier in RuntimeTransaction<SanitizedVersionedTransaction>::try_from() and never makes it into a batch

---

This is an automatic backport of pull request #9155 done by Mergify.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 83.2%. Comparing base (201f977) to head (110d979).

Additional details and impacted files

@@            Coverage Diff            @@
##             v3.1    #9848     +/-   ##
=========================================
- Coverage    83.2%    83.2%   -0.1%     
=========================================
  Files         865      865             
  Lines      376114   376085     -29     
=========================================
- Hits       313140   313083     -57     
- Misses      62974    63002     +28     

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[2501babe]

plan is to land this, rebase #9849, then rebase #9756