An update about WOFF 2.0 in Chrome

Hi working group members and WOFF 2.0 fans!

I’m pleased to share an update about WOFF 2.0 in Chrome.

As you most likely know, Chrome has supported WOFF 2.0 for almost a year
now. We initially used LZMA before switching to Brotli. In parallel, Google
Fonts has been serving WOFF 2.0 font files for the dynamic sub-setting
requests (i.e. requests using the text parameter).

On the security aspect, Chrome’s implementation and by extension the
reference implementation, passed the security review after our ClusterFuzz
infrastructure detected a couple of issues that were promptly fixed in
early February. The ClusterFuzz instance for WOFF 2.0 is still running and
as yet to find any additional issues since then.

In terms of feature quality, with WOFF 2.0 support in both Chrome and
Google Fonts, we’ve been able to test WOFF 2.0 end to end and confirmed
that everything was working fine.

Now, we are eager to see how Chrome’s WOFF 2.0 implementation and Google
Fonts’ WOFF 2.0 integration scale up as well as gather feedback and metrics
from real-word usage. In order to do so, we are tentatively
enabling<https://groups.google.com/a/chromium.org/d/msg/chromium-dev/j27Ou4RtvQI/PnvFh9XB-AkJ>WOFF
2.0 support by default in Chrome.

Clearly, there is still a lot of work in front of us but thanks to the hard
work and dedication of this working group, WOFF 2.0 has come a really long
way and we now have a reasonably optimal (the technical term is incredibly
good ;) proposal. We are eager to learn more from real world usage and
share our findings.