REXX SIGNAL ON ERROR Deployment Log Triage Failure Policy

A REXX deployment-log triage script uses SIGNAL ON ERROR to trap runtime errors and record diagnostics instead of terminating immediately. The goal is to decide the failure policy for trapped errors: abort the script on the first error, or log each error and continue through the deployment log to produce a complete failure summary.

Uncertainty remains around RC and SIGL semantics in the target interpreter, required filesystem privileges for reading deployment logs and writing the diagnostic report, and whether log lines follow a parseable format. Trapping a REXX runtime error also does not establish that an external command succeeded.

Should the triage script abort on the first trapped error, or continue after logging it? Which interpreter-specific condition functions are safe to rely on? What minimum privileges must the REXX runtime hold for log access and report output?