Warn when 'baseCommand' is a single string with whitespace (#2240)

A CommandLineTool 'baseCommand' given as a single string with internal
whitespace (e.g. "tar xf" instead of [tar, xf]) is passed to the OS as one
program name, which almost never exists. --validate previously accepted it
silently. Emit a warning at construction time pointing at the offending line
and suggesting the list form. List- and single-word baseCommands are
unaffected.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Author: arimu1

cwltool/command_line_tool.py

@@ -403,6 +403,17 @@ class CommandLineTool(Process):
     def __init__(self, toolpath_object: CommentedMap, loadingContext: LoadingContext) -> None:
         """Initialize this CommandLineTool."""
         super().__init__(toolpath_object, loadingContext)
+        base_command = self.tool.get("baseCommand")
+        if isinstance(base_command, str) and len(base_command.split()) > 1:
+            _logger.warning(
+                SourceLine(self.tool, "baseCommand", str).makeError(
+                    "'baseCommand' is a single string containing whitespace: %r. "
+                    "It will be passed to the operating system as one program name, "
+                    "which is unlikely to exist. If you meant a program plus arguments, "
+                    "provide a list instead, e.g. [%s]."
+                    % (base_command, ", ".join(repr(p) for p in base_command.split()))
+                )
+            )
         self.prov_obj = loadingContext.prov_obj
         self.path_check_mode: PathCheckingMode = (
             PathCheckingMode.RELAXED

tests/test_validate.py

@@ -4,6 +4,8 @@
 import logging
 import re
 
+import pytest
+
 from .util import get_data, get_main_output
 
 
@@ -125,3 +127,37 @@ def test_validate_custom_logger() -> None:
     assert "tests/CometAdapter.cwl#out' previously defined" not in stdout
     assert "tests/CometAdapter.cwl#out' previously defined" not in stderr
     assert "tests/CometAdapter.cwl#out' previously defined" in custom_log_text
+
+
+def test_validate_warns_on_basecommand_with_space() -> None:
+    """A 'baseCommand' string containing whitespace warns but still validates."""
+    custom_log = io.StringIO()
+    handler = logging.StreamHandler(custom_log)
+    handler.setLevel(logging.DEBUG)
+    exit_code, stdout, stderr = get_main_output(
+        ["--validate", get_data("tests/wf/2240-basecommand-space.cwl")],
+        logger_handler=handler,
+    )
+    custom_log_text = re.sub(r"\s\s+", " ", custom_log.getvalue())
+    assert exit_code == 0
+    assert "is valid CWL" in stdout
+    assert "'baseCommand' is a single string containing whitespace" in custom_log_text
+    assert "'tar xf'" in custom_log_text
+    assert "['tar', 'xf']" in custom_log_text
+
+
+@pytest.mark.parametrize(
+    "cwl_file",
+    ["tests/wf/2240-basecommand-list.cwl", "tests/CometAdapter.cwl"],
+)
+def test_validate_no_basecommand_space_warning(cwl_file: str) -> None:
+    """A list-form (or absent) 'baseCommand' does not trigger the whitespace warning."""
+    custom_log = io.StringIO()
+    handler = logging.StreamHandler(custom_log)
+    handler.setLevel(logging.DEBUG)
+    exit_code, _, _ = get_main_output(
+        ["--validate", get_data(cwl_file)],
+        logger_handler=handler,
+    )
+    assert exit_code == 0
+    assert "single string containing whitespace" not in custom_log.getvalue()

tests/wf/2240-basecommand-list.cwl

@@ -0,0 +1,6 @@
+#!/usr/bin/env cwl-runner
+cwlVersion: v1.2
+class: CommandLineTool
+baseCommand: [tar, xf]
+inputs: []
+outputs: []

tests/wf/2240-basecommand-space.cwl

@@ -0,0 +1,6 @@
+#!/usr/bin/env cwl-runner
+cwlVersion: v1.2
+class: CommandLineTool
+baseCommand: "tar xf"
+inputs: []
+outputs: []