2.1 COBOL copybook analysis

The analysis take place in 3 phases :

2.1.1 ''Cleaning'' the copybook

In this phase, the copybook is read and :

• comments are skipped,
• Multi-line COBOL declaration are concatenated in one line.
• PICTURE clauses are replaced by PIC,
• SIGN clauses are ignored,
• IS clauses are ignored,
• NATIVE clauses are skipped,
• X(4) and 9(6) symbols are replaced respectively by XXXX and 999999,
• Final dot of COBOL declarations is suppressed,
• COBOL declarations are capitalized.^2.1

2.1.2 Extracting the meta-data

In this phase, the following meta-data are extracted :

Meta-data

Usage COBOL level

FILLER clauses are not skipped. They are used to compute the starting positions of all COBOL items. They never appears in the XML file when the text to XML conversion process ends. The copybook is also checked for unsupported COBOL clauses :

• REDEFINES
• DEPENDING ON
• COMPUTATIONAL-1, COMPUTATIONAL-2, COMPUTATIONAL-5, COMP-1, COMP-2, and COMP-5.
• POINTER, PROCEDURE-POINTER and FUNCTION-POINTER
• OBJECT REFERENCE
• NATIONAL
• DISPLAY-1^2.2
• P symbol (decimal scaling position)
• INDEX
• level 66, 77, 88,^2.3
• Non-numeric COBOL levels or COBOL levels greater than 49,
• DBCS picture symbols.

and for partially supported COBOL clauses :

• OCCURS for elementary and group item with only one dimension,
• VALUES except for QUOTES and NULL clauses.

2.1.3 Renumbering the COBOL item levels

As COBOL item levels are used to indent tags during text to XML conversion, they are internally renumbered to have a smoother indentation. For example, the following COBOL copybooks :

Figure: A sample COBOL copybook

and

Figure: Same sample COBOL copybook with different levels

will always give the same indentation, like in the following XML file :

Figure: Resulting XML file after conversion

The original COBOL copybook is never modified by TXT2XML and if the conversion ends without errors, a report of this analysis is printed.