I can not reproduce this behavior. Isn't render preview a GPU texture? Not sure how this would cause such behavior. I would perhaps speculate, that while VSE cache is cleared, the VSE preview…
@mont29 Ah, indeed. Somehow I thought it is all tackled by manual's make update.
Anyway, seems an easy change to do.
It would indeed be good to expand those headers to RISC-V!
@mont29 I am not sure if there is anything actionable from the CI/CD perspective. We already have tracks for the different releases, so the changes are limited to the organization of the…
Not sure I understand here... I thought the problem was precisely that buildbots use the system python?
The problem is deeper ;) We have multiple different python version to worry about:
-…
Bring able to test things more automatically is surely great!
There are a few concerns with the current approach.
First one is the find_package(Python3 COMPONENTS Interpreter).
It will…
Overall looks good. The pixel offsets I'd leave to Omar to review.
It might be the safest approach for the 4.3.1, but it is not robust.
That doesn't feel like a proper English, and I don't really understand it. There is an extra comma which doesn't really belong there, and some weird double-negation.
The change seems fine, but it technically targets the wrong branch. The typical workflow after the release: