Explanation

The PDF Association announced yesterday that PDF 2.0 will get a new /BrotliDecode filter: https://pdfa.org/brotli-compression-coming-to-pdf/ The actual specification is still missing nevertheless, but they provide three example files of unknown license (although creating our own examples should work nevertheless).

This probably requires introducing a new dependency, for example on the reference implementation bindings: https://pypi.org/project/Brotli/ There might be pure Python implementations, but they lack active support and are much slower. Doing our own variant based upon RFC 7932 itself probably introduces too much maintenance overhead - while the actual complexity is not quite complex, it would involve much more code than for the existing filters (except the JBIG2 one which is not yet sorted out).