When Safe-Region Properties Reject Correct Repairs

A property proved on a region where a faulty program is safe need not hold on legitimate behavior outside that region. We examine this limitation in a bounded C repair-assessment pipeline. In the historical final study, all 23 admitted safety-passing repair calls matched the reference. Preservation accepted all 23, whereas a template-property gate flagged 4 and a trace-derived-property gate flagged 19. Detection of incorrect repairs therefore could not be estimated. A model-free continuation reuses these repairs and development pools, yielding 35 non-manufactured repair calls across 11 families. In six families, restricting a property check to its certified region eliminated false flags in at least one fully resolved comparison for a detector. This follows from preservation when the check uses only observations that preservation covers. Original and reference programs agreed on all 24 certified small and medium family-region pairs. The selection rule therefore left no contaminated certified region on which to assess harm. The audit documents abstentions, corrections to the analysis and a deviation from the solver timeout protocol.

Authors

Institutions

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-19
DOI
https://doi.org/10.5281/zenodo.24374637
Primary Topic
Software Testing and Debugging Techniques
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

When Safe-Region Properties Reject Correct Repairs

S. M. Abdulla Al Mamun
Zenodo (CERN European Organization for Nuclear Research)
Software Testing and Debugging Techniques
preprint

When Safe-Region Properties Reject Correct Repairs

S. M. Abdulla Al Mamun
preprint en

Abstract

A property proved on a region where a faulty program is safe need not hold on legitimate behavior outside that region. We examine this limitation in a bounded C repair-assessment pipeline. In the historical final study, all 23 admitted safety-passing repair calls matched the reference. Preservation accepted all 23, whereas a template-property gate flagged 4 and a trace-derived-property gate flagged 19. Detection of incorrect repairs therefore could not be estimated. A model-free continuation reuses these repairs and development pools, yielding 35 non-manufactured repair calls across 11 families. In six families, restricting a property check to its certified region eliminated false flags in at least one fully resolved comparison for a detector. This follows from preservation when the check uses only observations that preservation covers. Original and reference programs agreed on all 24 certified small and medium family-region pairs. The selection rule therefore left no contaminated certified region on which to assess harm. The audit documents abstentions, corrections to the analysis and a deviation from the solver timeout protocol.

Zenodo (CERN European Organization for Nuclear Research)
University of Manchester (GB)
Software Testing and Debugging Techniques
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon