Warn when XML subelement text exceeds lxml 2 GB limit

openmc/_xml.py

@@ -1,3 +1,31 @@
+_LXML_MAX_TEXT_SIZE = 2 * 1024**3  # 2 GB
+
+
+def _check_text_size(text, context=""):
+    """Raise an error if *text* exceeds the lxml 2 GB limit.
+
+    Parameters
+    ----------
+    text : str
+        Text that will be assigned to an XML element or attribute.
+    context : str
+        Human-readable description of what is being written, used in the
+        error message.
+
+    Raises
+    ------
+    ValueError
+        If len(text) >= 2 GB.
+    """
+    if len(text) >= _LXML_MAX_TEXT_SIZE:
+        size_gb = len(text) / 1024**3
+        raise ValueError(
+            f"The text content for {context!r} is ~{size_gb:.1f} GB, which "
+            f"exceeds the lxml limit of 2 GB. lxml will silently produce an "
+            f"empty element. Consider reducing the number of entries."
+        )
+
+
 def clean_indentation(element, level=0, spaces_per_level=2, trailing_indent=True):
     """Set indentation of XML element and its sub-elements.
     Copied and pasted from https://effbot.org/zone/element-lib.htm#prettyprint.

openmc/cell.py

@@ -8,7 +8,7 @@
 
 import openmc
 import openmc.checkvalue as cv
-from ._xml import get_elem_list, get_text
+from ._xml import _check_text_size, get_elem_list, get_text
 from .mixin import IDManagerMixin
 from .plots import add_plot_params
 from .region import Region, Complement
@@ -640,6 +640,7 @@ def create_xml_subelement(self, xml_element, memo=None):
             matlist_str = " ".join(
                 ["void" if m is None else str(m.id) for m in self.fill]
             )
+            _check_text_size(matlist_str, f"cell {self.id} material")
             material_subelement.text = matlist_str
 
         elif self.fill_type in ('universe', 'lattice'):
@@ -681,14 +682,18 @@ def create_surface_elements(node, element, memo=None):
         if self.temperature is not None:
             if isinstance(self.temperature, Iterable):
                 temperature_subelement= ET.SubElement(element, "temperature")
-                temperature_subelement.text = ' '.join(str(t) for t in self.temperature)
+                text = ' '.join(str(t) for t in self.temperature)
+                _check_text_size(text, f"cell {self.id} temperature")
+                temperature_subelement.text = text
             else:
                 element.set("temperature", str(self.temperature))
 
         if self.density is not None:
             if isinstance(self.density, Iterable):
                 density_subelement= ET.SubElement(element, "density")
-                density_subelement.text =  ' '.join(str(d) for d in self.density)
+                text = ' '.join(str(d) for d in self.density)
+                _check_text_size(text, f"cell {self.id} density")
+                density_subelement.text = text
             else:
                 element.set("density", str(self.density))
 

tests/unit_tests/test_cell.py

@@ -433,3 +433,9 @@ def test_plot(run_in_tmpdir):
     # ensure that calling the plot method doesn't
     # affect the universe ID space
     assert u_before.id + 1 == u_after.id
+
+
+def test_xml_text_size_check():
+    from openmc._xml import _check_text_size, _LXML_MAX_TEXT_SIZE
+    with pytest.raises(ValueError, match="exceeds the lxml limit"):
+        _check_text_size("x" * _LXML_MAX_TEXT_SIZE, "test element")