feat: add argument to specify the number of jobs being used for interfile

cli/src/semgrep/commands/scan.py

@@ -144,6 +144,10 @@ def convert(
  "--jobs",
  type=int,
  ),
+ optgroup.option(
+ "--interfile-jobs",
+ type=int,
+ ),
  optgroup.option(
  "--max-memory",
  type=int,
@@ -511,6 +515,7 @@ def scan(
  exclude_rule: Optional[Tuple[str, ...]],
  force_color: bool,
  include: Optional[Tuple[str, ...]],
+ interfile_jobs: Optional[int],
  jobs: Optional[int],
  lang: Optional[str],
  matching_explanations: bool,
@@ -719,6 +724,7 @@ def scan(
  try:
  metacheck_errors = CoreRunner(
  jobs=jobs,
+ interfile_jobs=interfile_jobs,
  engine_type=engine_type,
  timeout=timeout,
  max_memory=max_memory,
@@ -775,6 +781,7 @@ def scan(
  lang=lang,
  configs=(config or ["auto"]),
  no_rewrite_rule_ids=(not rewrite_rule_ids),
+ interfile_jobs=interfile_jobs,
  jobs=jobs,
  include=include,
  exclude={product: (exclude or ()) for product in ALL_PRODUCTS},

cli/src/semgrep/core_runner.py

@@ -476,6 +476,7 @@ class CoreRunner:
  def __init__(
  self,
  jobs: Optional[int],
+ interfile_jobs: Optional[int],
  engine_type: EngineType,
  timeout: int,
  max_memory: int,
@@ -489,6 +490,7 @@ def __init__(
  ):
  self._binary_path = engine_type.get_binary_path()
  self._jobs = jobs or engine_type.default_jobs
+ self._interfile_jobs = jobs or engine_type.default_interfile_jobs
  self._engine_type = engine_type
  self._timeout = timeout
  self._max_memory = max_memory
@@ -899,6 +901,7 @@ def _run_rules_direct_to_semgrep_core_helper(
  "-timeout_for_interfile_analysis",
  str(self._interfile_timeout),
  ]
+ cmd += ["-interfile_jobs", str(self._interfile_jobs)]
  cmd += [root]
  elif engine is EngineType.PRO_INTRAFILE:
  cmd += ["-deep_intra_file"]

cli/src/semgrep/engine.py

@@ -130,11 +130,13 @@ def get_cpu_count() -> int:
 
  @property
  def default_jobs(self) -> int:
- if self == EngineType.PRO_INTERFILE:
- return 1
  # Maxing out number of cores used to 16 if more not requested to not overload on large machines
  return min(16, self.get_cpu_count())
 
+ @property
+ def default_interfile_jobs(self) -> int:
+ return 1
+
  @property
  def default_max_memory(self) -> int:
  """Interfile uses a lot of memory, so it should have a safe default limit."""

cli/src/semgrep/run_scan.py

@@ -352,6 +352,7 @@ def run_scan(
  str
  ], # NOTE: Since the `ci` command reuses this function, we intentionally do not set a default at this level.
  no_rewrite_rule_ids: bool = False,
+ interfile_jobs: Optional[int] = None,
  jobs: Optional[int] = None,
  include: Optional[Sequence[str]] = None,
  exclude: Optional[Mapping[Product, Sequence[str]]] = None,
@@ -552,6 +553,7 @@ def run_scan(
  core_start_time = time.time()
  core_runner = CoreRunner(
  jobs=jobs,
+ interfile_jobs=interfile_jobs,
  engine_type=engine_type,
  timeout=timeout,
  max_memory=max_memory,