[lit] Fix 42812: lit test suite can no longer be run stand-alone

[stella.stamenova]

Summary

This change updates the lit.cfg file to use llvm_config when it is available, but when it is not, it directly modifies the config object. This makes it possible to run the lit tests standalone without having built llvm (as long as the correct binaries are present in the path such as FileCheck and not).

Because the lit tests don't take a hard dependency on llvm_config, some features such as system-windows have to have definitions in lit's cfg file as well. This is a potential issue as the os features sometimes change names (for example, we went from windows to system-windows, etc.). This can cause drift between lit's tests and the rest of the llvm tests.

[stella.stamenova]

@mgorny Could you download the patch and see if it fixes your scenario? I tested it on both Ubuntu and Linux, but neither installation of LLVM I have actually contains FileCheck or not, so my setup is not entirely the same as yours.

[edit] That would be both Ubuntu and Windows not Ubuntu and Linux. Duh.

[mgorny]

WFM.

llvm/trunk/utils/lit/tests/lit.cfg

 if llvm_src_root:
   # ``test_source_root`` may be in LLVM's binary build directory which does not contain
   # ``lit.py``, so use `llvm_src_root` instead.
   lit_path = os.path.join(llvm_src_root, 'utils', 'lit')
 else:
   lit_path = os.path.join(config.test_source_root, '..')
 
 # Required because some tests import the lit module
+if llvm_config:
   llvm_config.with_environment('PYTHONPATH', lit_path, append_path=True)
+else:
+  config.environment['PYTHONPATH'] = os.pathsep.join([lit_path])
 
 # Add llvm and lit tools directories if this config is being loaded indirectly.
+# In this case, we can also expect llvm_config to have been imported correctly.
 for attribute in ('llvm_tools_dir', 'lit_tools_dir'):
     directory = getattr(config, attribute, None)
     if directory:
         llvm_config.with_environment('PATH', directory, append_path=True)
 
 config.substitutions.append(('%{inputs}', os.path.join(
     config.test_source_root, 'Inputs')))
 config.substitutions.append(('%{lit}', "%%{python} %s" % (

 # Add a feature to detect if psutil is available
 supported, errormsg = lit_config.maxIndividualTestTimeIsSupported
 if supported:
     config.available_features.add("lit-max-individual-test-time")
 else:
     lit_config.warning('Setting a timeout per test not supported. ' + errormsg
                        + ' Some tests will be skipped and the --timeout'
                          ' command line argument will not work.')
+
+# When running the lit tests standalone, we want to define the same features
+# that the llvm_config defines. This means that the 'system-windows' feature
+# (and any others) need to match the names in llvm_config for consistency
+if not llvm_config:
+  if sys.platform.startswith('win') or sys.platform.startswith('cygwin'):
+    config.available_features.add('system-windows')