Charles Dang
Charles Dang
It probably would be if we bumped vcpkg, but it would be better to wait for the above patch so as to avoid any other download-from-gitlab issues.
https://github.com/microsoft/vcpkg/pull/48711 was merged. As soon as an action runner update lands with that commit we can bump our baseline and re-enable the Windows job.
Sure. First, the CMake issue, though.
*rubs eyes*. Let's get the master job up first, then ponder 1.18. Might just need to leave it broken... The current Windows action runner rollout (version 20251208.136.1) misses the above...
I think they're done weekly. That being said, we *could* bump now, since you reported no build issues, but I think better to wait than potentially having to bump again.
Ya know what, I'm just going to bump the baseline now and reenable the Windows job. We're going to raise the baseline again soon anyway to pull in SDL3, and...
The performance gains are great, and I really appreciate that you're chasing down inefficiencies in the render pipeline, but this PR is impossible to review as-is. It's a single commit...
Doc comments (and comments in general) are helpful when they convey information not otherwise conveyed elsewhere. They're a maintenance burden in and of themselves, so it's good to be judicious...
Can you split the halo::effect commit into its own PR? I'll merge that now.
Nice! Confirmed it fixes the issue above. Trying to figure out exactly *why* the fix works, though. Moving the update loop seems especially counterintuitive...