Clarify behavior for empty begin/while/repeat blocks in MASM (tests still ignored)

## Summary
There are still ignored tests for empty `begin end`, `while.true end`, and `repeat.N end` blocks, which suggests the intended behavior is unclear or not enforced. This leaves ambiguity for compiler/codegen output and users.

## Background
Historically, empty blocks were discussed and partially addressed:
- https://github.com/0xMiden/miden-vm/issues/585
- https://github.com/0xMiden/miden-vm/pull/609
- https://github.com/0xMiden/miden-vm/pull/1360

Despite that, empty `begin/while/repeat` appear to still be unresolved (tests remain ignored).

## Why this matters
- Code generators often emit empty blocks during optimization or transformation passes.
- Without a defined behavior, identical programs can be rejected or behave inconsistently across versions.
- It is unclear whether this should be allowed as a no-op or rejected with a structured diagnostic.

## Proposed resolution
Please confirm which behavior is desired:

Option A: Reject
- Emit a structured diagnostic for empty blocks.
- Enable the ignored tests and assert the diagnostics.

Option B: Allow
- Treat empty blocks as a no-op (e.g., a canonicalized `nop`).
- Enable tests to assert the resulting output/IR.
- Document the behavior in the assembly docs.

I’m happy to take this on once the intended behavior is confirmed.

Find the ignored tests for empty begin/while/repeat blocks. Look at the PRs #609 and #1360 to understand the history. Examine the assembly parser and codegen to see how these blocks are currently handled. Determine if the fix is to add a diagnostic or to treat them as a no-op, then update the tests accordingly.

由索引模型根据 Issue 内容生成。