Easytrieve Procedures Overview

Three Procedure Families

User PROCs behave like paragraphs or sections in other languages—explicit call, explicit return via END-PROC. Report and screen procedures are event hooks tied to product lifecycle. Mixing them confuses beginners who PERFORM BEFORE-LINE from JOB logic—that hook runs only when report engine prints a line.

Basic User PROC Syntax

1
2
3
4
5
6
7
8
9
10
11
12
13
CALC-TAX. PROC
  IF GROSS GT EXEMPT-AMT
     TAX = (GROSS - EXEMPT-AMT) * RATE
  ELSE
     TAX = 0
  END-IF
END-PROC
 
JOB INPUT PAYFILE
  PERFORM CALC-TAX
  IF TAX GT 0
     PERFORM WRITE-TAX-RECORD
  END-IF

Label ends with period before PROC keyword per Easytrieve conventions. Statements between PROC and END-PROC execute on each PERFORM. Control returns to statement after PERFORM unless EXIT or STOP inside PROC terminates broader flow.

PERFORM and Return Flow

PERFORM name transfers control once per invocation. Nested PERFORM stacks returns like subroutine calls. END-PROC is mandatory delimiter; omitting it merges procedures with catastrophic compile errors or runaway flow. GOTO inside PROC can branch within procedure or exit depending on target labels—prefer structured IF and PERFORM over GOTO for maintainability. STOP RUN inside PROC ends entire program—not just procedure.

Activity Scope Rules

JOB activity PROCs serve JOB INPUT and JOB logic. REPORT activity PROCs serve that report subactivity. SCREEN PROCs belong to named SCREEN block. SORT activity may define PROCs for sort processing hooks. Cannot PERFORM JOB PROC from REPORT PROC—duplicate shared code into a Library macro included in both activities if maintenance requires single source. Activity boundary is compile-time enforcement beginners hit when copying sample code across sections.

Report Procedure Placement

Broadcom requires report special procedures immediately after the REPORT subactivity they serve—not grouped at file bottom far from REPORT statement. BEFORE-LINE runs before detail line print; AFTER-LINE after. BEFORE-BREAK and AFTER-BREAK bracket control field changes. Sequence matters for readability and some release diagnostics. See report control breaks tutorial for WHEN BREAK interaction.

1
2
3
4
5
6
7
8
9
REPORT PAYRPT
  SEQUENCE DEPT EMP-NO
  LINE DEPT EMP-NO GROSS TAX
 
BEFORE-LINE. PROC
  IF LINE-COUNT GT 55
     SKIP 2
  END-IF
END-PROC

Screen Procedure Overview

INITIATION once at SCREEN start. BEFORE-SCREEN each display cycle before map send. AFTER-SCREEN after terminal input. TERMINATION at SCREEN end. Restrictions apply: GOTO SCREEN invalid in BEFORE-SCREEN. Screen procedures parallel report hooks but target online I/O. Batch programmers learning SCREEN should read screen section tutorial after this overview.

Nesting and Modular Design

PROC A may PERFORM B; B may PERFORM C. Depth limited by product and shop standards—not infinite recursion. A calling B calling A is recursion and forbidden. Design leaf PROCs for smallest reusable units—format date, validate state code—and composite PROCs that PERFORM leaves. Array iteration tutorials use PERFORM PROCESS-ELEMENT inside DO loops—same modular idea.

Procedures Versus Paragraphs

Easytrieve also documents paragraph labels for GOTO targets without full PROC envelope in some contexts. Procedures are preferred for PERFORM modularization. Paragraphs suit sparse GOTO flow in legacy code. New development should default to PROC and END-PROC for clarity. Paragraphs page compares details.

Procedures Versus Macros

MACRO in Library expands at compile time—copy source in multiple activities. PROC is runtime call in one activity. Share cross-activity logic with macros; share repeated logic within activity with PROC. Macro page covers WHEN to choose expansion versus call overhead—usually negligible in batch.

Section Learning Path

1. Procedures overview (this page) — map categories and scope.
2. PROC — label, PROC keyword, body rules.
3. END-PROC — termination and return semantics.
4. Nested procedures — call chains and depth.
5. Calling procedures — PERFORM variants and placement.
6. Procedure scope — activity and name visibility.
7. Procedure parameters — passing data when release supports parameter lists.

When to Create a Procedure

• Same logic appears twice in one activity.
• Detail and exception paths share validation block.
• Report line setup exceeds five lines—move to BEFORE-LINE PROC.
• Screen field initialization repeats each cycle—BEFORE-SCREEN PROC.
• Unit testing mindset: PROC boundary is logical test chunk in walkthroughs.

Common Beginner Mistakes

• PERFORM across activity boundary.
• Missing END-PROC.
• PERFORM report hook from JOB expecting manual invocation.
• Recursion between two PROCs cleaning up shared flags.
• STOP RUN inside utility PROC killing batch unexpectedly.

Explain It Like I'm Five

A procedure is a recipe card with a name. When the main program says perform recipe CALC-TAX, the computer follows that card and comes back to continue the main recipe. Some recipe cards are called automatically by the report printer before each line—that is BEFORE-LINE. You cannot use the report card from the homework section and the homework card from the report section—they live in different binders called activities.

Exercises

1. Write CALC-TAX PROC with END-PROC and one PERFORM from JOB INPUT.
2. List three special-name report procedures and when they run.
3. Explain why JOB PROC cannot be PERFORMed from REPORT.
4. Sketch call chain A performs B performs C without recursion.
5. Decide PROC versus MACRO for logic shared by JOB and REPORT.