ci(l1): reenable engine-cancun hive simulation

.github/scripts/check-hive-results.sh

@@ -10,6 +10,21 @@ if ! command -v jq >/dev/null 2>&1; then
   exit 1
 fi
 
+if ! command -v python3 >/dev/null 2>&1; then
+  echo "python3 is required to process Hive client logs but was not found in PATH"
+  exit 1
+fi
+
+slugify() {
+  local input="${1:-}"
+  local lowered trimmed
+  lowered="$(printf '%s' "${input}" | tr '[:upper:]' '[:lower:]')"
+  trimmed="$(printf '%s' "${lowered}" | sed -E 's/[^a-z0-9._-]+/-/g')"
+  trimmed="${trimmed#-}"
+  trimmed="${trimmed%-}"
+  printf '%s' "${trimmed}"
+}
+
 results_dir="${1:-src/results}"
 
 if [ ! -d "$results_dir" ]; then
@@ -86,15 +101,13 @@ for json_file in "${json_files[@]}"; do
       {
         echo "### Hive failures: ${suite_name:-$(basename "${json_file}" .json)}"
         printf '%s\n' "${failure_list}"
+        echo "Note: Hive scenarios may include multiple ethrex clients, so each failing case can have more than one log snippet."
         echo
       } >> "${GITHUB_STEP_SUMMARY}"
     fi
 
     suite_slug_raw="${suite_name:-$(basename "${json_file}" .json)}"
-    suite_slug="$(printf '%s' "${suite_slug_raw}" | tr '[:upper:]' '[:lower:]')"
-    suite_slug="$(printf '%s' "${suite_slug}" | sed -E 's/[^a-z0-9._-]+/-/g')"
-    suite_slug="${suite_slug#-}"
-    suite_slug="${suite_slug%-}"
+    suite_slug="$(slugify "${suite_slug_raw}")"
     suite_dir="${failed_logs_root}/${suite_slug:-suite}"
     mkdir -p "${suite_dir}"
 
@@ -178,6 +191,198 @@ for json_file in "${json_files[@]}"; do
       done <<< "${suite_logs_output}"
     fi
 
+    client_case_entries="$(
+      jq -r '
+        .testCases
+        | to_entries[]
+        | select(.value.summaryResult.pass != true)
+        | . as $case_entry
+        | ($case_entry.value.clientInfo? // {}) | to_entries[]
+        | [
+            .value.logFile // "",
+            ($case_entry.value.name // ("case-" + $case_entry.key)),
+            $case_entry.key,
+            ($case_entry.value.start // ""),
+            ($case_entry.value.end // ""),
+            .key
+          ]
+        | @tsv
+      ' "${json_file}" 2>/dev/null || true
+    )"
+    generated_client_snippets=0
+    if [ -n "${client_case_entries}" ]; then
+      client_logs_dir="${suite_dir}/client_logs"
+      mkdir -p "${client_logs_dir}"
+
+      while IFS= read -r client_entry; do
+        [ -n "${client_entry}" ] || continue
+        IFS=$'\t' read -r client_log_rel raw_case_name case_id case_start case_end client_id <<< "${client_entry}"
+
+        if [ -z "${client_log_rel}" ] || [ -z "${case_start}" ] || [ -z "${case_end}" ]; then
+          continue
+        fi
+
+        log_copy_path="${suite_dir}/${client_log_rel}"
+        if [ ! -f "${log_copy_path}" ]; then
+          continue
+        fi
+
+        case_slug="$(slugify "${raw_case_name}")"
+        if [ -n "${case_slug}" ]; then
+          case_slug="${case_slug}-case-${case_id}"
+        else
+          case_slug="case-${case_id}"
+        fi
+
+        client_slug="$(slugify "${client_id}")"
+        if [ -z "${client_slug}" ]; then
+          client_slug="client"
+        fi
+
+        case_dir="${client_logs_dir}/${case_slug}"
+        mkdir -p "${case_dir}"
+        snippet_path="${case_dir}/${client_slug}.log"
+
+        python3 - "${log_copy_path}" "${snippet_path}" "${raw_case_name}" "${case_start}" "${case_end}" "${client_id}" "${client_log_rel}" <<'PY'
+import sys
+from datetime import datetime, timedelta
+from pathlib import Path
+
+FORMATS = ("%Y-%m-%dT%H:%M:%S.%fZ", "%Y-%m-%dT%H:%M:%SZ")
+CONTEXT_SECONDS = 2
+PREFETCH_LIMIT = 50
+
+def normalise_timestamp_str(value):
+    if not value or not value.endswith("Z"):
+        return value
+    prefix = value[:-1]
+    if "." not in prefix:
+        return value
+    base, frac = prefix.split(".", 1)
+    frac_digits = "".join(ch for ch in frac if ch.isdigit())
+    if not frac_digits:
+        return f"{base}.000000Z"
+    frac_digits = (frac_digits + "000000")[:6]
+    return f"{base}.{frac_digits}Z"
+
+def parse_timestamp(value):
+    if not value:
+        return None
+    value = normalise_timestamp_str(value)
+    for fmt in FORMATS:
+        try:
+            return datetime.strptime(value, fmt)
+        except ValueError:
+            continue
+    return None
+
+def timestamp_from_line(line):
+    if not line:
+        return None
+    token = line.split(" ", 1)[0]
+    if not token or not token[0].isdigit():
+        return None
+    token = normalise_timestamp_str(token)
+    for fmt in FORMATS:
+        try:
+            return datetime.strptime(token, fmt)
+        except ValueError:
+            continue
+    return None
+
+log_path = Path(sys.argv[1])
+output_path = Path(sys.argv[2])
+case_name = sys.argv[3]
+case_start_raw = sys.argv[4]
+case_end_raw = sys.argv[5]
+client_id = sys.argv[6] or "unknown"
+client_log_rel = sys.argv[7]
+
+try:
+    log_content = log_path.read_text(encoding="utf-8", errors="replace").splitlines(keepends=True)
+except Exception as exc:
+    output_path.parent.mkdir(parents=True, exist_ok=True)
+    output_path.write_text(f"# Failed to read log '{log_path}': {exc}\n", encoding="utf-8")
+    sys.exit(0)
+
+start_ts = parse_timestamp(case_start_raw)
+end_ts = parse_timestamp(case_end_raw)
+
+fallback_reason = None
+if not start_ts or not end_ts or end_ts < start_ts:
+    fallback_reason = "Unable to determine reliable time window from test metadata."
+else:
+    start_ts = start_ts - timedelta(seconds=CONTEXT_SECONDS)
+    end_ts = end_ts + timedelta(seconds=CONTEXT_SECONDS)
+
+captured_lines = []
+prefetch = []
+current_ts = None
+capturing = False
+
+if not fallback_reason:
+    for line in log_content:
+        ts = timestamp_from_line(line)
+        if ts is not None:
+            current_ts = ts
+
+        if not capturing:
+            prefetch.append(line)
+            if len(prefetch) > PREFETCH_LIMIT:
+                prefetch.pop(0)
+
+        in_window = current_ts is not None and start_ts <= current_ts <= end_ts
+
+        if in_window:
+            if not capturing:
+                captured_lines.extend(prefetch)
+                capturing = True
+            captured_lines.append(line)
+        elif capturing and current_ts is not None and current_ts > end_ts:
+            break
+        elif capturing:
+            captured_lines.append(line)
+
+    if not captured_lines:
+        fallback_reason = "No timestamped log lines matched the computed time window."
+
+if fallback_reason:
+    captured_lines = log_content
+
+header_lines = [
+    f"# Test: {case_name}\n",
+    f"# Client ID: {client_id}\n",
+    f"# Source log: {client_log_rel}\n",
+]
+
+if start_ts and end_ts and not fallback_reason:
+    header_lines.append(
+        f"# Time window (UTC): {case_start_raw} .. {case_end_raw} (with ±{CONTEXT_SECONDS}s context)\n"
+    )
+else:
+    header_lines.append("# Time window (UTC): unavailable\n")
+
+if fallback_reason:
+    header_lines.append(f"# NOTE: {fallback_reason}\n")
+
+header_lines.append("\n")
+
+output_path.parent.mkdir(parents=True, exist_ok=True)
+with output_path.open("w", encoding="utf-8") as dst:
+    dst.writelines(header_lines)
+    dst.writelines(captured_lines)
+PY
+
+        if [ -s "${snippet_path}" ]; then
+          generated_client_snippets=$((generated_client_snippets + 1))
+        fi
+      done <<< "${client_case_entries}"
+    fi
+
+    if [ "${generated_client_snippets}" -gt 0 ]; then
+      echo "Generated ${generated_client_snippets} client log snippet(s) in ${client_logs_dir}"
+    fi
+
     echo "Saved Hive failure artifacts to ${suite_dir}"
 
     failures=$((failures + failed_cases))

.github/workflows/pr-main_l1.yaml

@@ -146,55 +146,35 @@ jobs:
             simulation: ethereum/rpc-compat
             # https://github.com/ethereum/execution-apis/pull/627 changed the simulation to use a pre-merge genesis block, so we need to pin to a commit before that
             buildarg: "branch=d08382ae5c808680e976fce4b73f4ba91647199b"
-            hive_repository: ethereum/hive
-            hive_version: 7709e5892146c793307da072e1593f48039a7e4b
             artifact_prefix: rpc_compat
           - name: "Devp2p tests"
             simulation: devp2p
             limit: discv4|eth|snap/Ping|Findnode/WithoutEndpointProof|Findnode/PastExpiration|Amplification|Status|StorageRanges|ByteCodes|GetBlockHeaders|SimultaneousRequests|SameRequestID|ZeroRequestID|GetBlockBodies|MaliciousHandshake|MaliciousStatus|NewPooledTxs|GetBlockReceipts|BlockRangeUpdate|GetTrieNodes
             # Findnode/BasicFindnode fails due to packets being processed out of order
             # Findnode/UnsolicitedNeighbors flaky in CI very occasionally. When fixed replace all "Findnode/<test>" with "Findnode"
-            hive_repository: ethereum/hive
-            hive_version: 7709e5892146c793307da072e1593f48039a7e4b
             artifact_prefix: devp2p
           - name: "Engine Auth and EC tests"
             simulation: ethereum/engine
             limit: engine-(auth|exchange-capabilities)/
-            hive_repository: ethereum/hive
-            hive_version: 7709e5892146c793307da072e1593f48039a7e4b
             artifact_prefix: engine_auth_ec
-          # - name: "Cancun Engine tests"
-          #   simulation: ethereum/engine
-          #   limit: "engine-cancun"
-          #   hive_repository: ethereum/hive
-          #   hive_version: 7709e5892146c793307da072e1593f48039a7e4b
-          #   artifact_prefix: engine_cancun
+          - name: "Cancun Engine tests"
+            simulation: ethereum/engine
+            limit: "engine-cancun"
+            artifact_prefix: engine_cancun
           - name: "Paris Engine tests"
             simulation: ethereum/engine
             limit: "engine-api"
-            hive_repository: ethereum/hive
-            hive_version: 7709e5892146c793307da072e1593f48039a7e4b
             artifact_prefix: engine_paris
           - name: "Engine withdrawal tests"
             simulation: ethereum/engine
             limit: "engine-withdrawals/Corrupted Block Hash Payload|Empty Withdrawals|engine-withdrawals test loader|GetPayloadBodies|GetPayloadV2 Block Value|Max Initcode Size|Sync after 2 blocks - Withdrawals on Genesis|Withdraw many accounts|Withdraw to a single account|Withdraw to two accounts|Withdraw zero amount|Withdraw many accounts|Withdrawals Fork on Block 1 - 1 Block Re-Org|Withdrawals Fork on Block 1 - 8 Block Re-Org NewPayload|Withdrawals Fork on Block 2|Withdrawals Fork on Block 3|Withdrawals Fork on Block 8 - 10 Block Re-Org NewPayload|Withdrawals Fork on Canonical Block 8 / Side Block 7 - 10 Block Re-Org [^S]|Withdrawals Fork on Canonical Block 8 / Side Block 9 - 10 Block Re-Org [^S]"
-            hive_repository: ethereum/hive
-            hive_version: 7709e5892146c793307da072e1593f48039a7e4b
             artifact_prefix: engine_withdrawals
           # Investigate this test
           # - name: "Sync"
           #   simulation: ethereum/sync
           #   limit: ""
-          #   hive_repository: ethereum/hive
-          #   hive_version: 7709e5892146c793307da072e1593f48039a7e4b
           #   artifact_prefix: sync
     steps:
-      - name: Free Disk Space (Ubuntu)
-        uses: jlumbroso/free-disk-space@v1.3.1
-        with:
-          tool-cache: false
-          large-packages: false
-
       - name: Checkout sources
         uses: actions/checkout@v4
 
@@ -225,7 +205,7 @@ jobs:
           SIM_LIMIT: ${{ matrix.limit }}
           SIM_BUILDARG: ${{ matrix.buildarg }}
         run: |
-          FLAGS='--sim.parallelism 16 --sim.loglevel 1'
+          FLAGS='--sim.parallelism 4 --sim.loglevel 3'
           if [[ -n "$SIM_LIMIT" ]]; then
             escaped_limit=${SIM_LIMIT//\'/\'\\\'\'}
             FLAGS+=" --sim.limit '$escaped_limit'"
@@ -239,8 +219,8 @@ jobs:
         id: run-hive-action
         uses: ethpandaops/hive-github-action@v0.5.0
         with:
-          hive_repository: ${{ matrix.hive_repository }}
-          hive_version: ${{ matrix.hive_version }}
+          hive_repository: ethereum/hive
+          hive_version: 7709e5892146c793307da072e1593f48039a7e4b
           simulator: ${{ matrix.simulation }}
           client: ethrex
           client_config: ${{ steps.client-config.outputs.config }}