feat(scripts): use dynamic addresses in ported static tests (#2695)

Convert all ported static tests under `tests/ported_static/` from hard-coded addresses (literal `Address(0x…)` and `EOA(key=…)`) to dynamic allocation (`pre.fund_eoa()`, `pre.deploy_contract()` without `address=`). Tests now behave like hand-written EEST tests, where the framework picks addresses at fill time. 68 of 2151 tests become runnable under the `execute` plugin against live networks (no `pre_alloc_mutable` marker).

Achieved without changing test semantics — verified by 39,002 / 39,002 trace comparisons equivalent under `exact-no-stack` against pre-conversion baseline traces.

### What changed

**`scripts/filler_to_python/analyzer.py`** — bulk of the work. New analyzer logic:

- Topological sort of contracts by bytecode address references; substitute known addresses with variable names (`addr_to_var`) in emitted Op expressions.
- Resolve CREATE-derived addresses in post-state storage and account references via `compute_create_address(...)` (depth 2, nonce 0–15).
- Coinbase pinning: any account whose address equals `env.current_coinbase` is kept hardcoded so the env's `fee_recipient` and the pre-state entry agree.
- Sender pinning when post-state has unresolvable addresses or address-like storage values (>2³² ints) — keeps CREATE addresses derivable from sender consistent with baseline.
- Short-PUSH detection: contracts referenced via `PUSH<20` (i.e. addresses with leading-zero bytes) are pinned to baseline addresses.
- Same for EOAs (precompile-touch case): EOAs referenced via short PUSH or hint-overridden into the precompile range (`<eoa:0x...01>`) are pinned literally at `0x01–0x10`.
- Computed call targets (`address=Op.ADD/MLOAD/CALLDATALOAD/...`) and address arithmetic trigger a global hardcoded fallback.
- `FORCE_HARDCODED_TESTS` allowlist (144 entries) for tests that fundamentally don't survive dynamic addresses (gas-precise tests, keccak-derived storage keys, cryptographic signatures in tx-data, CREATE2 collisions, structural pre-execution rejections).
- `needs_mutable_pre` flag — emitted only when `assert_mutable()` would actually fire (sender or any account hardcoded, sender funded with non-default nonce, or any contract emitted with `nonce=0`).

**`scripts/filler_to_python/ir.py`** — `use_dynamic` on `AccountIR`/`SenderIR`/`AccessListEntryIR`, `nonce` on `SenderIR`, `needs_mutable_pre` on `IntermediateTestModel`.

**`scripts/filler_to_python/templates/state_test.py.j2`** — conditional rendering: `pre.fund_eoa(amount=...)` / `pre.deploy_contract(...)` for dynamic tests, `EOA(key=...)` / `pre[var] = Account(...)` / `pre.deploy_contract(..., address=Address(...))` for hardcoded ones. `@pytest.mark.pre_alloc_mutable` is only emitted when `needs_mutable_pre`.

**`scripts/filler_to_python/render.py`** — `format_storage` handles string variable references; new context fields piped to the template.

**`tests/ported_static/`** — all 2151 test files regenerated. 14 `@manually-enhanced` files preserved as-is via the existing skip mechanism.

### Verification

| Check | Result |
|---|---|
| Regeneration | 2137 / 2151 OK + 14 skipped (manually-enhanced) |
| `just static` | clean |
| Full fill (CI mode) | 60,476 / 60,476 passed |
| Full fill `--traces` | 60,476 / 60,476 passed (depends on PR #2709 in `forks/amsterdam`) |
| `--verify-traces` `exact-no-stack` | **39,002 / 39,002 equivalent** |

> **Note:** The trace report shows 78 additional "divergences" on 3 STRUCTURAL tests (CREATE-collision / EIP-3607). These are spurious and tracked in #2758 — an asymmetry between in-memory and disk-loaded traces in the verify-traces machinery (extends #2709). The tests themselves pass, fixtures are byte-identical to baseline, and the comparison artifact is independent of this PR.

Author: leolara

scripts/filler_to_python/analyzer.py

@@ -36,6 +36,7 @@ class AccountIR:
     code_expr: str = ""
     storage: dict = field(default_factory=dict)
     oversized_code: bool = False
+    use_dynamic: bool = True
 
 
 @dataclass
@@ -80,6 +81,7 @@ class AccessListEntryIR:
 
     address: str = ""
     storage_keys: list = field(default_factory=list)
+    use_dynamic: bool = False
 
 
 @dataclass
@@ -108,7 +110,9 @@ class SenderIR:
     is_tagged: bool = False
     key: int | None = None
     balance: int = 0
+    nonce: int | None = None
     not_in_pre: bool = False
+    use_dynamic: bool = True
 
 
 @dataclass
@@ -136,6 +140,7 @@ class IntermediateTestModel:
     is_slow: bool = False
     is_multi_case: bool = False
     is_fork_dependent: bool = False
+    needs_mutable_pre: bool = False
     environment: EnvironmentIR = field(
         default_factory=lambda: EnvironmentIR(
             coinbase_var="coinbase", number=0, timestamp=0

scripts/filler_to_python/ir.py

@@ -42,7 +42,11 @@ def format_storage(d: dict) -> str:
         return "{}"
     items = []
     for k in sorted(d.keys()):
-        items.append(f"{format_int(k)}: {format_int(d[k])}")
+        v = d[k]
+        if isinstance(v, str):
+            items.append(f"{format_int(k)}: {v}")
+        else:
+            items.append(f"{format_int(k)}: {format_int(v)}")
     single = "{" + ", ".join(items) + "}"
     if len(single) <= 50:
         return single
@@ -271,6 +275,7 @@ def render_test(ir: IntermediateTestModel) -> str:
         "is_slow": ir.is_slow,
         "is_multi_case": ir.is_multi_case,
         "is_fork_dependent": ir.is_fork_dependent,
+        "needs_mutable_pre": ir.needs_mutable_pre,
         "has_exceptions": has_exceptions,
         "env": ir.environment,
         "accounts": ir.accounts,

scripts/filler_to_python/render.py

@@ -159,7 +159,9 @@ def _storage_with_any(base: dict, any_keys: list) -> Storage:
 {% if has_exceptions and not is_multi_case %}
 @pytest.mark.exception_test
 {% endif %}
+{% if needs_mutable_pre %}
 @pytest.mark.pre_alloc_mutable
+{% endif %}
 def {{ test_name }}(
     state_test: StateTestFiller,
     pre: Alloc,
@@ -176,7 +178,13 @@ def {{ test_name }}(
 {% for addr in address_constants %}
     {{ addr.var_name }} = Address({{ addr.hex }})
 {% endfor %}
-{% if sender.not_in_pre %}
+{% if sender.use_dynamic %}
+{% if sender.nonce %}
+    sender = pre.fund_eoa(amount={{ sender.balance | format_int }}, nonce={{ sender.nonce }})
+{% else %}
+    sender = pre.fund_eoa(amount={{ sender.balance | format_int }})
+{% endif %}
+{% elif sender.not_in_pre %}
     sender = pre.fund_eoa(amount=0)
 {% else %}
     sender = EOA(
@@ -205,15 +213,19 @@ def {{ test_name }}(
 
 {# Pre-state account setup #}
 {% for account in accounts %}
-{% if account.is_sender %}
+{% if account.is_sender and sender.use_dynamic %}
+{# Sender balance already set via pre.fund_eoa() above #}
+{% elif account.is_sender %}
     pre[sender] = Account(balance={{ account.balance | format_int }}{{ ", nonce=%d" | format(account.nonce) if account.nonce }}{{ ", storage=%s" | format(account.storage | format_storage) if account.storage }}{{ ", code=%s" | format(account.code_expr | wrap_op_chain(indent=8)) if account.code_expr }})
+{% elif account.is_eoa and account.use_dynamic %}
+    {{ account.var_name }} = pre.fund_eoa(amount={{ account.balance | format_int }})  # noqa: F841
 {% elif account.is_eoa %}
     pre[{{ account.var_name }}] = Account(balance={{ account.balance | format_int }}{{ ", nonce=%d" | format(account.nonce) if account.nonce }}{{ ", storage=%s" | format(account.storage | format_storage) if account.storage }}{{ ", code=%s" | format(account.code_expr | wrap_op_chain(indent=8)) if account.code_expr }})
 {% else %}
 {# Contract: source comment + deploy #}
 {{ account.source_comment }}
 {% if account.oversized_code %}
-    {{ account.var_name }} = Address({{ account.address }})
+    {{ account.var_name }} = Address({{ account.address }})  # oversized contract
     pre[{{ account.var_name }}] = Account(
         code={{ account.code_expr | wrap_op_chain(indent=8) }},
 {% if account.storage %}
@@ -238,7 +250,9 @@ def {{ test_name }}(
 {% if account.nonce is not none %}
         nonce={{ account.nonce }},
 {% endif %}
+{% if not account.use_dynamic %}
         address=Address({{ account.address }}),  # noqa: E501
+{% endif %}
     )
 {% endif %}
 {% endif %}
@@ -287,7 +301,11 @@ def {{ test_name }}(
         {{ d_idx }}: [
     {% for al in al_entries %}
             AccessList(
+    {% if al.use_dynamic %}
+                address={{ al.address }},
+    {% else %}
                 address=Address({{ al.address }}),
+    {% endif %}
                 storage_keys=[
     {% for sk in al.storage_keys %}
                     Hash("{{ sk }}"),  # noqa: E501
@@ -357,7 +375,11 @@ def {{ test_name }}(
         access_list=[
 {% for al in tx.access_list %}
             AccessList(
+{% if al.use_dynamic %}
+                address={{ al.address }},
+{% else %}
                 address=Address({{ al.address }}),
+{% endif %}
                 storage_keys=[
 {% for sk in al.storage_keys %}
                     Hash(

scripts/filler_to_python/templates/state_test.py.j2

@@ -0,0 +1,34 @@
+#!/bin/bash
+# Verify that filler_to_python with dynamic addresses produces
+# trace-equivalent tests. Assumes output/traces_baseline/ already
+# exists (generated once before any code changes).
+set -euo pipefail
+
+export TMPDIR=./.tmp
+mkdir -p "$TMPDIR" output/traces_new
+
+if [ ! -d "output/traces_baseline" ]; then
+    echo "ERROR: output/traces_baseline/ not found."
+    echo "Generate baseline first (before code changes):"
+    echo "  TMPDIR=./.tmp uv run fill tests/ported_static/ --evm-dump-dir output/traces_baseline -n 10 -m 'not slow'"
+    exit 1
+fi
+
+# Step 1: Run filler_to_python (overwrites tests/ported_static/)
+echo "=== Step 1: Running filler_to_python ==="
+uv run python -m scripts.filler_to_python \
+    --fillers tests/static/static/state_tests/ \
+    --output tests/ported_static/
+
+# Step 2: Fill new tests + verify against baseline
+echo "=== Step 2: Filling new tests and verifying traces ==="
+uv run fill \
+    tests/ported_static/ \
+    --evm-dump-dir output/traces_new \
+    --verify-traces output/traces_baseline \
+    --verify-traces-comparator exact-no-stack \
+    -n 10 \
+    -m "not slow"
+
+echo "=== Done. Check output above for trace mismatches ==="
+echo "=== Use 'git diff tests/ported_static/' to see code changes ==="

scripts/verify_dynamic_addresses.sh

@@ -7,7 +7,6 @@
 
 import pytest
 from execution_testing import (
-    EOA,
     Account,
     Address,
     Alloc,
@@ -58,9 +57,7 @@ def test_add_non_const(
 ) -> None:
     """Test_add_non_const."""
     coinbase = Address(0x2ADC25665018AA1FE0E6BC666DAC8FC2697FF9BA)
-    sender = EOA(
-        key=0xB1F4CBC3A50042184425A6F9E996D0910F7BA879457CE5DAC5C71E498AD3C005
-    )
+    sender = pre.fund_eoa(amount=0xDE0B6B3A7640000)
 
     env = Environment(
         fee_recipient=coinbase,
@@ -85,7 +82,6 @@ def test_add_non_const(
         nonce=0,
         address=Address(0xF1722FE346FA35E045DE07E47CF6AF9BAE8ADE0A),  # noqa: E501
     )
-    pre[sender] = Account(balance=0xDE0B6B3A7640000)
 
     expect_entries_: list[dict] = [
         {

tests/ported_static/stArgsZeroOneBalance/test_add_non_const.py

@@ -7,7 +7,6 @@
 
 import pytest
 from execution_testing import (
-    EOA,
     Account,
     Address,
     Alloc,
@@ -58,9 +57,7 @@ def test_addmod_non_const(
 ) -> None:
     """Test_addmod_non_const."""
     coinbase = Address(0x2ADC25665018AA1FE0E6BC666DAC8FC2697FF9BA)
-    sender = EOA(
-        key=0xB1F4CBC3A50042184425A6F9E996D0910F7BA879457CE5DAC5C71E498AD3C005
-    )
+    sender = pre.fund_eoa(amount=0xDE0B6B3A7640000)
 
     env = Environment(
         fee_recipient=coinbase,
@@ -86,7 +83,6 @@ def test_addmod_non_const(
         nonce=0,
         address=Address(0x92D2FC80312ACD8C37857696D2224AF18CE6F966),  # noqa: E501
     )
-    pre[sender] = Account(balance=0xDE0B6B3A7640000)
 
     expect_entries_: list[dict] = [
         {

tests/ported_static/stArgsZeroOneBalance/test_addmod_non_const.py

@@ -7,7 +7,6 @@
 
 import pytest
 from execution_testing import (
-    EOA,
     Account,
     Address,
     Alloc,
@@ -58,9 +57,7 @@ def test_and_non_const(
 ) -> None:
     """Test_and_non_const."""
     coinbase = Address(0x2ADC25665018AA1FE0E6BC666DAC8FC2697FF9BA)
-    sender = EOA(
-        key=0xB1F4CBC3A50042184425A6F9E996D0910F7BA879457CE5DAC5C71E498AD3C005
-    )
+    sender = pre.fund_eoa(amount=0xDE0B6B3A7640000)
 
     env = Environment(
         fee_recipient=coinbase,
@@ -85,7 +82,6 @@ def test_and_non_const(
         nonce=0,
         address=Address(0x4C26357E0D164B702BCEB18690FC742EE1D36913),  # noqa: E501
     )
-    pre[sender] = Account(balance=0xDE0B6B3A7640000)
 
     expect_entries_: list[dict] = [
         {

tests/ported_static/stArgsZeroOneBalance/test_and_non_const.py

@@ -7,7 +7,6 @@
 
 import pytest
 from execution_testing import (
-    EOA,
     Account,
     Address,
     Alloc,
@@ -58,9 +57,7 @@ def test_balance_non_const(
 ) -> None:
     """Test_balance_non_const."""
     coinbase = Address(0x2ADC25665018AA1FE0E6BC666DAC8FC2697FF9BA)
-    sender = EOA(
-        key=0xB1F4CBC3A50042184425A6F9E996D0910F7BA879457CE5DAC5C71E498AD3C005
-    )
+    sender = pre.fund_eoa(amount=0xDE0B6B3A7640000)
 
     env = Environment(
         fee_recipient=coinbase,
@@ -86,7 +83,6 @@ def test_balance_non_const(
         nonce=0,
         address=Address(0xEE6A324B2ECE5EACDF881ABFDCC62B5361D0FB50),  # noqa: E501
     )
-    pre[sender] = Account(balance=0xDE0B6B3A7640000)
 
     expect_entries_: list[dict] = [
         {

tests/ported_static/stArgsZeroOneBalance/test_balance_non_const.py

@@ -7,7 +7,6 @@
 
 import pytest
 from execution_testing import (
-    EOA,
     Account,
     Address,
     Alloc,
@@ -58,9 +57,7 @@ def test_byte_non_const(
 ) -> None:
     """Test_byte_non_const."""
     coinbase = Address(0x2ADC25665018AA1FE0E6BC666DAC8FC2697FF9BA)
-    sender = EOA(
-        key=0xB1F4CBC3A50042184425A6F9E996D0910F7BA879457CE5DAC5C71E498AD3C005
-    )
+    sender = pre.fund_eoa(amount=0xDE0B6B3A7640000)
 
     env = Environment(
         fee_recipient=coinbase,
@@ -85,7 +82,6 @@ def test_byte_non_const(
         nonce=0,
         address=Address(0x86D606901085BA78C64D2E0B16831E6AFD89DE2D),  # noqa: E501
     )
-    pre[sender] = Account(balance=0xDE0B6B3A7640000)
 
     expect_entries_: list[dict] = [
         {