Replies: 1 comment 1 reply
-
I agree that this area could be improved. Our artifacts experience remains experimental today, and unfortunately, we don't currently have the capacity to address this, but I hope we can revisit it soon. |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Heyo!
Happy thanksgiving for those that celebrate! 🦃
Recently I tried to start using vcpkg artifacts (
vcpkg activate
) with an existing repository which uses vcpkg in manifest mode as a submodule. This didn't work!What I eventually realized after reading the blog post more clearly is that
vcpkg-init.cmd
played a critical part in providing a wrapper around vcpkg which will modify the environment and so on. Makes sense.BUT
That does present a certain adoption friction over our existing submodule (and fetch content) approach.
What's the plan here? What are we doing? Having vcpkg as a submodule is really incredible and our cmake presets accomplish a great deal as a result. Currently our toolchains for embedded targets (and some associated tools for flashing hardware) get airlifted in via magic, but it would appear that vckpg artifacts are the more forward facing or long term approach to this issue.
In an ideal world, once a build is configured, it wouldn't be dependent on someone running
vcpkg-init.cmd
first especially when I consider how to integrate this into developer workflows with vscode or clion etc. Currently we lean heavily on cmake presets to setup the environment based on apriori knowledge of what's going to show up before the first call toproject()
in our build.What would really be the sauce is that some mechanism in the vcpkg toolchain would also initialize the environment based on artifact dependencies.
Beta Was this translation helpful? Give feedback.
All reactions