Merge branch 'main' into TAT-1043-feat-add-perps-slippage-config

Author: abretonc7s

.github/scripts/benchmark-stats-commit.sh

@@ -122,15 +122,21 @@ assemble_performance_data() {
 
     # Merge the startup group into presets (only if any startup files were found).
     if [[ "${page_load_json}" != "{}" ]]; then
-        presets_json=$(echo "${presets_json}" | jq --argjson pl "${page_load_json}" '. + {"pageLoad": $pl}')
+        # Keep large JSON off argv (ARG_MAX): presets_json via stdin, page_load_json
+        # via a temp file read with --slurpfile.
+        pl_file="$(mktemp)"
+        printf '%s' "${page_load_json}" >"${pl_file}"
+        presets_json=$(printf '%s' "${presets_json}" | jq --slurpfile pl "${pl_file}" '. + {"pageLoad": $pl[0]}')
+        rm -f "${pl_file}"
     fi
 
     echo "Collected ${file_count} preset(s)" >&2
 
-    jq -n \
+    # presets_json can exceed ARG_MAX; pass it via stdin instead of as a jq argument
+    # (a too-large argv makes the kernel fail to exec jq with "Argument list too long").
+    printf '%s' "${presets_json}" | jq \
         --argjson timestamp "$(date +%s000)" \
-        --argjson presets "${presets_json}" \
-        '{ timestamp: $timestamp, presets: $presets }'
+        '{ timestamp: $timestamp, presets: . }'
 }
 
 # Resolve stats file and assemble data
@@ -191,8 +197,14 @@ fi
 
 TEMP_FILE="${STATS_FILE}.tmp"
 
-jq --arg sha "${HEAD_COMMIT_HASH}" --argjson data "${COMMIT_DATA}" \
-    '. + {($sha): $data}' "${STATS_FILE}" > "${TEMP_FILE}"
+# COMMIT_DATA wraps presets_json and is strictly larger, so it can exceed
+# ARG_MAX too. Read it via --slurpfile from a temp file rather than passing
+# the blob on argv.
+data_file="$(mktemp)"
+printf '%s' "${COMMIT_DATA}" >"${data_file}"
+jq --arg sha "${HEAD_COMMIT_HASH}" --slurpfile data "${data_file}" \
+    '. + {($sha): $data[0]}' "${STATS_FILE}" > "${TEMP_FILE}"
+rm -f "${data_file}"
 mv "${TEMP_FILE}" "${STATS_FILE}"
 
 git add "${STATS_FILE}"