Fix a numpy test for 32 bit archs

The default dtype used in numpy array is "int". This type has
different sizes on 32 and 64 bit architectures. The test tries to use
the numpy array as a Long64_t dataframe. On 32 bit architectures this
results in an error.

This commit explicitly changes the numpy array's dtype to "int64" to
match the Long64_t dataframe type on all architectyres.

(cherry picked from commit 2ab39c1)

Author: ellert

bindings/pyroot/pythonizations/test/rdataframe_misc.py

@@ -132,7 +132,7 @@ def test_regression_gh_20291(self):
             return
         out_path = "dataframe_misc_regression_gh20291.root"
         try:
-            x, y = numpy.array([1, 2, 3]), numpy.array([4, 5, 6])
+            x, y = numpy.array([1, 2, 3], dtype='int64'), numpy.array([4, 5, 6], dtype='int64')
             df = ROOT.RDF.FromNumpy({"x": x, "y": y})
 
             df.Snapshot("tree", out_path)