diff options
| author |  Mike Klein <mtklein@chromium.org> | 2017-07-21 13:11:37 -0400 |
|---|---|---|
| committer |  Skia Commit-Bot <skia-commit-bot@chromium.org> | 2017-07-21 17:47:24 +0000 |
| commit | ce98801bbb5ac251b374f8b565445add43017aaa (patch) | |
| tree | 28cb9dbdc430844c78962b12d64353d3d82f77c6 /tests/PremulAlphaRoundTripTest.cpp | |
| parent | 7634330c98356548cacd63e805688d20579fd1f5 (diff) | |
when encoding PNG, don't assume F16 values are in range
I believe this will fix this batch of broken colorImages:
e23e55ef33358b2f6f98fc9cb795c71397c01618&fdiffmax=-1&fref=false&frgbamax=255&frgbamin=0&head=true&include=false&limit=50&match=gamma_correct&match=name&metric=combined&neg=false&offset=0&pos=false&query=source_type%3DcolorImage&sort=desc&unt=true">https://gold.skia.org/search?blame=e23e55ef33358b2f6f98fc9cb795c71397c01618&fdiffmax=-1&fref=false&frgbamax=255&frgbamin=0&head=true&include=false&limit=50&match=gamma_correct&match=name&metric=combined&neg=false&offset=0&pos=false&query=source_type%3DcolorImage&sort=desc&unt=true
Depending on the instruction set, clamping to logical [0,1] may be
implicit or not as we work our way down to bytes. I think that's why
these broken images are not more widespread on the bots.
Change-Id: Ie9bb937864bf6954301d76d9921a2d3029d12c9a
Reviewed-on: https://skia-review.googlesource.com/25742
Reviewed-by: Leon Scroggins <scroggo@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>
Diffstat (limited to 'tests/PremulAlphaRoundTripTest.cpp')
0 files changed, 0 insertions, 0 deletions