Can't build with CUDA-Q when project already uses a version of `fmt`

Required prerequisites

• Consult the security policy. If reporting a security vulnerability, do not report the bug using this form. Use the process described in the policy to report the issue.
• Make sure you've read the documentation. Your issue may be addressed there.
• Search the issue tracker to verify that this hasn't already been reported. +1 or comment there if it has.
• If possible, make a PR with a failing test to give us a starting point to work on!

Describe the bug

This is sort of related to https://github.com/NVIDIA/cuda-quantum/issues/1704. When building my project which uses CUDA-Q, I get a linking error because the version of fmt lib used in CUDA-Q is far older (3 years old commit) than the version used in my project. I don't think the version of fmt used in a downstream project should be constrained by the version used by CUDA-Q.

Steps to reproduce the bug

1. Create a cmake project that uses fmt lib and CUDA-Q
2. Use a recent version of fmt lib in the cmake project (e.g. 11.1.4)
3. Create a hello-world source file which uses:
   1. fmt::print() and fmt::join()
   2. some cuda-q functions that also use fmt
4. Try to build

You will get an error saying there are ambiguous definitions of fmt functions. I.e. cuda-q has linked against the older version of fmt and the example project has linked against the newer version.

Expected behavior

It builds.

Is this a regression? If it is, put the last known working version (or commit) here.

Not a regression

Environment

• CUDA-Q version: latest (0.11.0)
• Python version: n/a
• C++ compiler: g++
• Operating system: ubuntu 22.04

Suggestions

Either

1. upgrade the version of fmt in cuda-q
2. allow me to specify the version of fmt i want to use with the build