Implement end-to-end and regression tests for Markdown and MathML rendering

[fibonatto]

Problem

There are no automated tests for the Markdown or LaTeX/MathML parsing and rendering logic. This poses risks:

• Undetected regressions on code changes/refactoring
• No safety net for contributors
• Hard to verify behavior across edge cases (e.g., unclosed blocks, large posts, exotic Unicode)
• Lower trust in deployment quality

Proposed Solution

• Add a test harness (in C or JS, as appropriate) to render representative Markdown/Math and assert the output HTML/MathML is as expected
• Include tests for both happy paths and edge/error cases (nested headers, invalid syntax, very large post, malicious strings)
• Consider including Golden Files for output comparison
• Add at least one CI job to run regression tests on each PR

Acceptance Criteria

• Tests fail on parsing or rendering regressions
• All parsing logic is covered by at least one automated test
• Documentation instructs contributors how to run tests locally

References

• https://github.com/commonmark/cmark (test approach)
• https://github.com/google/googletest