removed dup/drop bindings

[TomerStarkware]

Remove dup/drop overrides; rely on arena ownership

Closes #NA

Now that all heap-allocated types (Box<T>, Nullable<T>, Array<T>, dicts, and the structs/enums containing them) live in the per-invocation
arena, deep clones and recursive drops are no longer necessary. Memory is reclaimed when the arena is destroyed at invocation end, and pointer-typed
values can be shared freely between branches because no libfunc mutates through an input box pointer.

This PR removes the DupOverridesMeta / DropOverridesMeta metadata and all associated code:

• Deletes src/metadata/drop_overrides.rs and src/metadata/dup_overrides.rs (~420 lines).
• dup libfunc becomes a bitwise copy — return the same SSA value twice.
• drop libfunc becomes a no-op.
• Per-type dup/drop builders are removed from array, box, enum, nullable, snapshot, and struct (~1300 lines).
• array_pop / array_get no longer deep-copy popped elements; they return GEP'd pointers into the source arena slot.
• array_slice, array_len, circuit_eval, get_circuit_output, sha256_process_block, and debug::print drop their now-redundant drop
  invocations on consumed inputs.
• struct_boxed_deconstruct / enum_boxed_match return field/payload pointers via GEP instead of unboxing and re-boxing each member.

Net diff: ~+300/-2100 in src/.

Introduces Breaking Changes?

No. The runtime ABI and libfunc semantics are unchanged — dup still returns two values, drop still consumes one, and consumed boxed values still
cannot be reused by the Sierra program. Only the MLIR codegen behind these libfuncs changes.

• Created PR in sequencer
• Created PR in starknet-replay
• Updated the starknet-blocks.yml workflow to use these PRs.

---

This change is 

[orizi]

@orizi reviewed 20 files and all commit messages, and made 2 comments.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on TomerStarkware).

---

src/libfuncs/const.rs line 77 at r1 (raw file):

    let inner_layout = const_ty.layout(registry)?;

    // TODO: For struct/enum const types we could emit the constant directly into the arena slot

can't this point to some constant memory?
i'm assuming llvm has some such feature.

[orizi]

@orizi reviewed 22 files and all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on TomerStarkware).

[TomerStarkware]

@TomerStarkware made 1 comment.
Reviewable status: 23 of 39 files reviewed, 1 unresolved discussion (waiting on orizi).

---

src/libfuncs/const.rs line 77 at r1 (raw file):

Previously, orizi wrote…

can't this point to some constant memory?
i'm assuming llvm has some such feature.

Done.

[orizi]

@orizi reviewed 17 files and all commit messages, made 1 comment, and resolved 1 discussion.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on TomerStarkware).

[github-actions]

Benchmark results Main vs HEAD.

Base

Command	Mean [s]	Min [s]	Max [s]	Relative
base dict_insert.cairo (JIT)	1.904 ± 0.023	1.869	1.943	1.03 ± 0.02
base dict_insert.cairo (AOT)	1.852 ± 0.023	1.821	1.887	1.00

Head

Command	Mean [s]	Min [s]	Max [s]	Relative
head dict_insert.cairo (JIT)	1.856 ± 0.020	1.816	1.888	1.02 ± 0.02
head dict_insert.cairo (AOT)	1.829 ± 0.019	1.805	1.856	1.00

Base

Command	Mean [s]	Min [s]	Max [s]	Relative
base dict_snapshot.cairo (JIT)	1.758 ± 0.020	1.724	1.780	1.06 ± 0.01
base dict_snapshot.cairo (AOT)	1.652 ± 0.012	1.638	1.677	1.00

Head

Command	Mean [s]	Min [s]	Max [s]	Relative
head dict_snapshot.cairo (JIT)	1.666 ± 0.013	1.647	1.686	1.01 ± 0.01
head dict_snapshot.cairo (AOT)	1.644 ± 0.013	1.628	1.660	1.00

Base

Command	Mean [s]	Min [s]	Max [s]	Relative
base factorial_2M.cairo (JIT)	2.195 ± 0.018	2.175	2.227	1.02 ± 0.01
base factorial_2M.cairo (AOT)	2.144 ± 0.007	2.132	2.154	1.00

Head

Command	Mean [s]	Min [s]	Max [s]	Relative
head factorial_2M.cairo (JIT)	2.157 ± 0.020	2.121	2.191	1.00 ± 0.01
head factorial_2M.cairo (AOT)	2.152 ± 0.018	2.124	2.183	1.00

Base

Command	Mean [s]	Min [s]	Max [s]	Relative
base fib_2M.cairo (JIT)	1.710 ± 0.009	1.700	1.721	1.03 ± 0.01
base fib_2M.cairo (AOT)	1.655 ± 0.009	1.642	1.667	1.00

Head

Command	Mean [s]	Min [s]	Max [s]	Relative
head fib_2M.cairo (JIT)	1.676 ± 0.014	1.657	1.700	1.01 ± 0.01
head fib_2M.cairo (AOT)	1.656 ± 0.016	1.632	1.680	1.00

Base

Command	Mean [s]	Min [s]	Max [s]	Relative
base linear_search.cairo (JIT)	1.794 ± 0.011	1.774	1.813	1.06 ± 0.02
base linear_search.cairo (AOT)	1.699 ± 0.024	1.671	1.752	1.00

Head

Command	Mean [s]	Min [s]	Max [s]	Relative
head linear_search.cairo (JIT)	1.714 ± 0.018	1.691	1.747	1.00
head linear_search.cairo (AOT)	1.724 ± 0.022	1.687	1.756	1.01 ± 0.02

Base

Command	Mean [s]	Min [s]	Max [s]	Relative
base logistic_map.cairo (JIT)	2.009 ± 0.017	1.983	2.038	1.10 ± 0.01
base logistic_map.cairo (AOT)	1.825 ± 0.012	1.805	1.840	1.00

Head

Command	Mean [s]	Min [s]	Max [s]	Relative
head logistic_map.cairo (JIT)	2.008 ± 0.035	1.956	2.080	1.09 ± 0.02
head logistic_map.cairo (AOT)	1.845 ± 0.022	1.821	1.889	1.00

[github-actions]

Benchmarking results

Benchmark for program dict_insert

Open benchmarks

Command	Mean [s]	Min [s]	Max [s]	Relative
Cairo-vm (Rust, Cairo 1)	11.505 ± 0.096	11.381	11.687	6.07 ± 0.10
cairo-native (embedded AOT)	1.914 ± 0.032	1.863	1.966	1.01 ± 0.02
cairo-native (embedded JIT using LLVM's ORC Engine)	1.896 ± 0.027	1.861	1.938	1.00

Benchmark for program dict_snapshot

Open benchmarks

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
Cairo-vm (Rust, Cairo 1)	557.9 ± 6.0	552.0	570.2	1.00
cairo-native (embedded AOT)	1734.7 ± 28.4	1677.3	1775.0	3.11 ± 0.06
cairo-native (embedded JIT using LLVM's ORC Engine)	1762.7 ± 29.1	1715.8	1802.7	3.16 ± 0.06

Benchmark for program factorial_2M

Open benchmarks

Command	Mean [s]	Min [s]	Max [s]	Relative
Cairo-vm (Rust, Cairo 1)	4.992 ± 0.021	4.964	5.039	2.25 ± 0.03
cairo-native (embedded AOT)	2.226 ± 0.026	2.176	2.267	1.00 ± 0.02
cairo-native (embedded JIT using LLVM's ORC Engine)	2.216 ± 0.025	2.181	2.266	1.00

Benchmark for program fib_2M

Open benchmarks

Command	Mean [s]	Min [s]	Max [s]	Relative
Cairo-vm (Rust, Cairo 1)	4.914 ± 0.023	4.887	4.954	2.82 ± 0.05
cairo-native (embedded AOT)	1.742 ± 0.031	1.706	1.795	1.00
cairo-native (embedded JIT using LLVM's ORC Engine)	1.754 ± 0.026	1.702	1.794	1.01 ± 0.02

Benchmark for program linear_search

Open benchmarks

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
Cairo-vm (Rust, Cairo 1)	612.4 ± 12.5	593.4	635.3	1.00
cairo-native (embedded AOT)	1757.3 ± 36.3	1712.2	1805.2	2.87 ± 0.08
cairo-native (embedded JIT using LLVM's ORC Engine)	1784.9 ± 23.1	1754.9	1830.9	2.91 ± 0.07

Benchmark for program logistic_map

Open benchmarks

Command	Mean [ms]	Min [ms]	Max [ms]	Relative
Cairo-vm (Rust, Cairo 1)	503.9 ± 13.1	494.0	539.6	1.00
cairo-native (embedded AOT)	1864.3 ± 17.9	1844.9	1895.6	3.70 ± 0.10
cairo-native (embedded JIT using LLVM's ORC Engine)	1987.9 ± 21.1	1963.2	2019.2	3.95 ± 0.11

[orizi]

@orizi reviewed 1 file and all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on TomerStarkware).