W3C MathML 3 Testsuite

This repository contains a local mirror of the W3C MathML 3 Full Testsuite, scraped from https://www.w3.org/Math/testsuite/build/main/overview.html .

Source

• URL: https://www.w3.org/Math/testsuite/build/main/overview.html

• Base URL: https://www.w3.org/Math/testsuite/build/

• Original testsuite runner: http://i2geo.net/mml-testsuite/main/

Why this repo exists

The W3C provides a downloadable testsuite for MathML v2, but no such archive exists for MathML v3. This repo provides a scraped copy so the MathML v3 testsuite is locally available. A similar scraper would be needed for any future MathML v4 testsuite.

Contents

Directory/Category	Description
Characters	Character/block tests (Blocks, Symbols, Variants, EntityNames, NumericRefs, UTF8)
Content	Content MathML element tests (ArithmeticAlgebraLogic, BasicContentElements, Calculus, etc.)
StrictContent	Strict Content MathML tests
Presentation	Presentation MathML element tests (CSS, GeneralLayout, ScriptsAndLimits, TablesAndMatrices, TokenElements)
General	General tests (Clipboard, GenAttribs, Math)
ErrorHandling	Error handling tests (BadAttribs, BadChildren, BadEntities, BadTags, MathML1mtable, NumChildren)
Topics	Topic-specific tests (BiDi, LineBreak, StretchyChars, Accents, Nesting)
TortureTests	Large/complex expression tests

File Types

• .mml — Raw MathML source (primary test data)

• .xhtml — Test page in 5 view variants: full, full5, simple, plain, form, slideshow

• .png / .gif — Reference rendering images (some tests share these)

Known Issues

• 43 URLs in the original testsuite return 404 (missing full5 variants, some malformed links, etc.). These are broken on the W3C server and cannot be downloaded.

• The scraper (scrape.py) is idempotent — re-running it only downloads missing files.

License

This testsuite is made available by the copyright holders under the W3C Test Suite License (2008 version). The testsuite was created circa 2012; per that license you are permitted to copy and distribute the contents for any purpose and without fee or royalty, provided you include the following on all copies you use:

1. A link or URL to the original document:
   https://www.w3.org/Math/testsuite/build/main/overview.html

2. The pre-existing copyright notice (verbatim from the license):
   

   Copyright © 2012 World Wide Web Consortium, (MIT, ERCIM, Keio, Beihang) and others. All Rights Reserved. https://www.w3.org/copyright/test-suites-licenses/

3. Attribution to the original author where it appears in the test files (e.g. Author: David Carlisle, per each test file).

The tests themselves shall NOT be changed. The W3C name and trademarks may not be used in advertising without prior written permission.

This work is provided AS IS without warranty; see the license for details.

Scraping

pip install -r requirements.txt

# Run from the repo root
python3 scripts/scrape.py