huge refactor for high RewritePatterns usage and less ad-hoc cpp code

remove Spatial many ops in favor of tensor ops like in pim

validation/subprocess_utils.py

@@ -47,7 +47,9 @@ def _stream_output(fd, process, reporter, treat_eio_as_eof=False, stream_output=
     return_code = process.wait()
     if return_code != 0:
         error_output = captured_output if not stream_output else recent_output
-        raise subprocess.CalledProcessError(return_code, process.args, output=bytes(error_output))
+        exc = subprocess.CalledProcessError(return_code, process.args, output=bytes(error_output))
+        exc.output_already_streamed = stream_output and bool(captured_output)
+        raise exc
     return bytes(captured_output)
 
 
@@ -67,15 +69,15 @@ def run_command_with_reporter(cmd, cwd=None, reporter=None, capture_output=False
 
     stream_output = bool(getattr(reporter, "verbose", False))
     if not stream_output:
-        process = subprocess.Popen(
+        completed = subprocess.run(
             cmd,
             cwd=cwd,
             stdout=subprocess.PIPE,
             stderr=subprocess.STDOUT,
         )
-        assert process.stdout is not None
-        output = _stream_output(process.stdout.fileno(), process, reporter, stream_output=False)
-        return output.decode("utf-8", errors="replace") if capture_output else None
+        if completed.returncode != 0:
+            raise subprocess.CalledProcessError(completed.returncode, completed.args, output=completed.stdout)
+        return completed.stdout.decode("utf-8", errors="replace") if capture_output else None
 
     try:
         master_fd, slave_fd = pty.openpty()