feat: multi-language code parsing and XML support (v3.6.0)

[lyonzin]

Summary

• Multi-language code parsing: C (.c), C++ (.cpp/.h), JavaScript (.js/.jsx), TypeScript (.ts/.tsx) with per-language regex extraction of functions, classes/structs/interfaces, imports/includes
• XML parser (.xml) with root element and namespace metadata extraction
• All 8 new formats default enabled — no config change needed
• Code parser now reports correct language metadata per file type (was hardcoded to "python" for all code files including .mqh/.mq4)
• Version bump 3.5.2 → 3.6.0

Changes

File	What changed
mcp_server/ingestion.py	LANGUAGE_PROFILES constant, language-aware _parse_code, new _parse_xml, dispatch table updated
mcp_server/config.py	_SUPPORTED_SUFFIXES + supported_formats default + fallback list
config.example.yaml	New extensions added with comments
presets/developer.yaml	All code formats enabled
presets/cybersecurity.yaml	C/C++/XML enabled
tests/conftest.py	5 new fixtures: sample_c, sample_cpp, sample_js, sample_ts, sample_xml
tests/test_ingestion.py	12 new tests including Python regression guard
tests/test_config.py	2 new tests for format validation
README.md	Format table 12→20, changelog v3.6.0
pyproject.toml	Version 3.6.0

Safety

• Zero regression: Python _parse_code uses character-identical regex patterns via LANGUAGE_PROFILES[".py"]
• Explicit regression guard: test_python_parse_unchanged validates Python output is identical
• All existing tests pass unmodified: 94/94
• Ruff lint + format: clean

Test plan

• All 94 tests pass (28 existing + 66 from other test files + new)
• Ruff lint clean
• Ruff format clean
• CI matrix (Python 3.11 + 3.12, Ubuntu + Windows)
• Release v3.6.0 to validate full pipeline (PyPI + NPM + Docker)