Luca

the original suggestion is to save the user to specify a value for `vcpkgGitCommitId` when this information is already stored in the `vcpkg.json` file (i.e. to avoid duplication). @equeim What...

@DanEngelbrecht this may be a documentation issue, something i'll take a look at soon. For your info, the action is optimizing the usage of the cache by using a primary...

@DanEngelbrecht if I got this right, the problem is that only for `macOS x86_64` workflow the cache is restored (i.e. cache-hit), but then when vcpkg is launched, the cached artifacts...

@DanEngelbrecht i think the title of this issue is misleading, since it is actually by design. What would be a better title that reflects the problem?

@DrMcCoy thanks for the report! Is this a persisting error? If so, you could verbosely listing the content of the cache during a workflow run, and see if that files...

@akrieger thanks for the detailed explanation. Regarding `prependedCacheKey`, that is something that certainly it can be added to `run-vcpkg`. Additional answers here below. > There are multiple negative side effects...

@akrieger IIRC there is a test on the commit id to represent a SHA1 string which would prevent your idea to work, which would be neat BTW. I think prepending...

FYI a`prependedKey` input is being worked in this PR https://github.com/lukka/run-vcpkg/pull/161

The latest version of `run-vcpkg@v10` adds the `prependedCacheKey` input, which helps in invalidating previously created caches. It also documents how to properly handle the saving of the cache when a...

@bigerl would be useful to create the missing tags for the CMake versions your scenario requires? Would this all you need? Or tags are not good enough for your scenario?...