which generally runs fine. However, when we update the compatibility of one of our dependencies (for example in this PR) the baseline benchmark fails with something like:
ERROR: LoadError: Unsatisfiable requirements detected for package NDTensors [23ae76d9]:
[...]
I suppose this is because it is trying to use the same version for the NDTensors dependency for both the target and baseline benchmark, but I don't know enough about BencharmCI.jl/PkgBenchmark.jl to understand why that is happening. Is there a way that I can adjust my configuration to get this case working? I tried explicitly setting the correct compat versions in benchmark/Project.toml for both the master branch and the PR branch but it gives the same error.
Thanks for your work on so many great packages!
We have a pretty simple setup for BenchmarkCI:
https://github.com/ITensor/ITensors.jl/tree/v0.1.41/benchmark https://github.com/ITensor/ITensors.jl/blob/v0.1.41/.github/workflows/benchmark.yml
which generally runs fine. However, when we update the compatibility of one of our dependencies (for example in this PR) the baseline benchmark fails with something like:
I suppose this is because it is trying to use the same version for the
NDTensorsdependency for both the target and baseline benchmark, but I don't know enough aboutBencharmCI.jl/PkgBenchmark.jlto understand why that is happening. Is there a way that I can adjust my configuration to get this case working? I tried explicitly setting the correctcompatversions inbenchmark/Project.tomlfor both themasterbranch and the PR branch but it gives the same error.