diffoutputdirs reports changes in WMO headers when other checks do not

[RussellManser-NCO]

What's wrong?

While upgrading petss/1.4.5, I found that diffoutputdirs reported changes in WMO headers between production and parallel runs. This was concerning for what was intended to be a minimal upgrade. Output from diffoutputdirs can be found at:

/lfs/h1/nco/idsb/noscrub/russell.manser/projects/petss_v1_4_5/diffoutputdirs/prod_para/diffoutputdirs.*etss.20260628.log

After some discussion with other team members and the developers, I compared WMO headers with the following short script.

#!/bin/bash

# Search for WMO headers in two files, then compare the results
#
# See the following: https://www.weather.gov/tg/headef

file1=$1
file2=$2

if [[ -z "${file1}" || -z "${file2}" ]]; then
  echo "Usage $0: file1 file2"
  exit 1
fi

diff <(strings "${file1}" | grep -E "[A-Z]{4}[0-9]{2}\ .*\ [0-9]{6}") \
<(strings "${file2}" | grep -E "[A-Z]{4}[0-9]{2}\ .*\ [0-9]{6}")

This script did not report any differences in WMO headers.

Proposed Solution

I suggest adding unit tests for diffoutputdirs to identify its intended behavior and locate any bugs. Once that is complete, I optionally suggest refactoring the script and improving its readability.