Tape/Z: An Overview

Avishek Sen Gupta on 22 June 2025

---

Abstract

We introduce Tape/Z, an evolving toolkit for analysing mainframe HLASM (High Level Assembler) code. The library provides capabilities for working with mainframe assembler code, including parsing, control flow graph building, dependency tracing, and flowchart visualization capabilities.

This post has not been written or edited by AI.

Writing the HLASM grammar

See Building an HLASM grammar from scratch for more details.

HLASM and SQL: Multi-pass parsing

Broadly, this extracts structure using multiple passes over the code. The purposes of these passes are detailed below:

• Extracts labels and determines their boundaries
• Concatenates lines which are marked as a single line spread over multiple lines (Preprocessing)
• Parses core HLASM using the generated grammar from the IBM spec
• Parses DB2 SQL using the grammar used in the Che4z for COBOL project (with slight changes)
• Sanitises code by removing extraneous characters after valid instructions (Preprocessing)
• Identifies unstructured macros (single-line macros)
• Identifies structured programming constructs (IF…ELSE…ENDIF)

The results are represented in a single JSON structure which contains HLASM, SQL, and macros, and is suitable for consumption into a graph for further analysis.

Control Flow Graph

References

• Tape/Z
• Cobol-REKT

---