I think it's time to make this update.
The 1.6 branch has NEON optimized filtering functions for Arm, improving decode performance by about 28% on average.
In addition, Skia has upstreamed SSE optimized filtering functions to 1.6.22, meaning that we see about a 34% performance improvement on Intel.
https://docs.google.com/a/google.com/spreadsheets/d/1Xv3TlgwDwnyOKdBrdzQeibz1oaeBQm6_cpPP0Wu0ou0/edit?usp=sharing
Skia has been testing with this code since January, and it will ship in Android N.
In terms of security, I don't think this makes much of a difference. libpng fixes security vulnerabilites as they arise for both 1.6 and 1.2. Both branches seem to be affected every once in a while.
http://www.libpng.org/pub/png/libpng.html
I'd be happy to take this on. I think it makes sense to wait until 1.6.22 is no longer in beta.
Comment 1 by noel@chromium.org
, Apr 1 2016