We have had the suggestion to revisit our cmake messages to understand when our printing is actually informative vs confusing.
Some examples:
- The hdf5 and h5py incomaptibility is a bold red error, although as I understand it, it just ditches pbarlike. So it should really not seem like a significant error.
- Repeated messages about missing backends / scanners for different versions of the same backend / scanner. If it found a version, it should be a problem.
- A lot of successful cmake messages are printed to error stream
- "Werror is disabled" is printed in red, suggesting it might be a problem.
We have had the suggestion to revisit our cmake messages to understand when our printing is actually informative vs confusing.
Some examples: