Regression in 1.8.2: XML root handling may be a str (not Element), causing crash in extract_enum_qualified_name #37
Description
Summary
After upgrading from coverxygen 1.8.1 to 1.8.2, our documentation coverage job crashes when extract_enum_qualified_name() tries to iterate the XML root.
This is a problem in 1.8.2 (1.8.1 works with the same inputs). The change in how the XML root is handled is not backward‑compatible and causes failures when a string is passed instead of an XML Element.
Environment
- Coverxygen: 1.8.2 (fails), 1.8.1 (works)
- Python: 3.10
- OS: Ubuntu 24.04
- Invocation: `python3 -m coverxygen --xml-dir <DOXYGEN_XML_DIR> --src-dir <SRC_DIR> --output - --format summary
Expected behavior
Coverxygen parses Doxygen XML and either:
- produces a coverage report (lcov/json/summary), or
- gracefully skips/excludes symbols per filters.
- No crash.
Minimal example to reproduce the model is attached.