strategy to reduce noise in generated output consumed by doc builds

Things like file system paths, ctest test numbers, runtimes, addresses (in some cases -- but some are important!) can contrib noise to the review process for examples.

See discussion here:

https://github.com/LLNL/conduit/pull/617

think about if/how we could clean the output to reduce noise.

Start with the discussion in PR #617 and inspect the generated example output consumed by the documentation builds, focusing on paths, ctest numbers, runtimes, and addresses. Define which values are noise versus important output and document a concrete approach whose effect can be reviewed in the generated examples.

Written by the indexing model from the issue text.