https://bugs.kde.org/show_bug.cgi?id=409414
Eoin O'Neill <eoinoneill1...@gmail.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|CONFIRMED |RESOLVED
Resolution|--- |FIXED
Latest Commit| |https://invent.kde.org/grap
| |hics/krita/commit/4a5eb9d53
| |78ac6041dcaa54ed82197923434
| |5f22
--- Comment #5 from Eoin O'Neill <eoinoneill1...@gmail.com> ---
Git commit 4a5eb9d5378ac6041dcaa54ed821979234345f22 by Eoin O'Neill.
Committed on 25/08/2020 at 04:30.
Pushed by eoinoneill into branch 'master'.
`pixelDataAtTime` Implementation Correction
An incorrect call to KisRasterKeyframeChannel::fetchFrame was causing undefined
behavior
when reading from a specific frame. Instead of passing in a write-target paint
device,
it was using the existing node's paint device which was causing the frame at
current time
to be overwritten by essentially arbitrary frame data.
This user-error is easy to make with the original method name ('fetchFrame'
doesn't sound
like a method that writes to a paint device, which made argument 2 somewhat
confusing
from an outsiders perspective) and has already been addressed in our Refactor,
but it's
probably worth fixing this first in an isolated commit.
Additional Note: the paint device being made to write to is using a snapshot
copy to create
the temporary device before feeding into the fetchFrame method. Since we
basically just
want an empty 'sheet' to write to, the contents of the frame should be mostly
arbitrary before
writing.
M +1 -1 libs/libkis/Node.cpp
https://invent.kde.org/graphics/krita/commit/4a5eb9d5378ac6041dcaa54ed821979234345f22
--
You are receiving this mail because:
You are watching all bug changes.
