Replace iterative capacity growth with next_power_of_two

[SaiCharanMarrivada]

Replace the capacity-doubling loop with (reserve_cap + len).next_power_of_two(). Given the existing invariant that capacity is a power of two, the result is identical.

[Copilot]

Pull request overview

This PR replaces an iterative capacity-doubling loop with a direct calculation using next_power_of_two() in the deque's reserve function. This is a straightforward optimization that maintains the same behavior while being more efficient and readable.

Changes:

• Replaced the capacity-doubling loop (lines 337-340 in the original code) with a single call to (reserve_cap + len).next_power_of_two()
• Updated the inline comment to reflect the new approach

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[SaiCharanMarrivada]

To ensure that replacing the iterative capacity growth with next_power_of_two() maintains exact semantic parity and handles edge cases correctly, I have verified the transformation using Alive2.

• LLVM-IR Generation: I’ve isolated the logic in Compiler Explorer to generate the IR for both the original loop and the new implementation.
• Alive2 Proof: You can view the formal equivalence proof here: Alive2 Verification.

Note on Verification Scope:
I used u32 for the proof to ensure the solver completes in a reasonable time.