JCL jobs, steps and DD statements

//DAILYPST JOB (ACCT),'DAILY POSTING',CLASS=A,MSGCLASS=X
//* Concept listing, not run: posts the day's transactions, then
//* prints the totals of each branch from what was posted.
//POST     EXEC PGM=POSTTXN
//STEPLIB  DD DISP=SHR,DSN=BANK.PROD.LOAD
//TXNIN    DD DISP=SHR,DSN=BANK.PROD.TXN.DAILY(0)
//POSTED   DD DSN=BANK.PROD.POSTED(+1),
//            DISP=(NEW,CATLG,DELETE),
//            SPACE=(TRK,(10,5),RLSE),
//            DCB=(RECFM=FB,LRECL=22)
//SYSOUT   DD SYSOUT=*
//REPORT   EXEC PGM=BRREPORT
//STEPLIB  DD DISP=SHR,DSN=BANK.PROD.LOAD
//POSTED   DD DISP=SHR,DSN=BANK.PROD.POSTED(+1)
//BRANCHES DD *
MUM MUMBAI
PUN PUNE
/*
//BRREPT   DD SYSOUT=*
//SYSOUT   DD SYSOUT=*

"""Reads the subset of JCL this lesson teaches and checks a deck before anyone submits it: the fields of every
statement (z/OS MVS JCL Reference, JCL statement fields), continuations, the steps and their DD statements, and
whether every ASSIGN name of each step's COBOL program has a DD statement. It runs nothing on z/OS.

Rules checked: // in columns 1 and 2; a name in column 3 of 1 to 8 letters, digits or # @ $, the first not a digit;
an operation of JOB, EXEC or DD; nothing past column 71; a statement whose parameters end with a comma continues on the
next line, which leaves column 3 blank. A DD with no name adds a data set to the DD before it (a concatenation).
"""
import re
import sys
from pathlib import Path

NAME = re.compile(r"[A-Z#@$][A-Z0-9#@$]{0,7}$")
ASSIGN = re.compile(r"ASSIGN\s+TO\s+\"?([A-Z0-9#@$-]+)\"?")
SYSTEM_DDS = {"STEPLIB", "SYSOUT", "SYSPRINT", "SYSUDUMP", "CEEDUMP"}


def read(path):
    problems, steps, statements, data_lines = [], [], 0, 0
    continuing, instream, seen_job = False, False, False
    for n, line in enumerate(Path(path).read_text().splitlines(), start=1):
        if line.startswith("//*"):
            continue
        if line.startswith("/*"):
            instream = False
            continue
        if not line.startswith("//"):
            if instream:
                data_lines += 1
            else:
                problems.append(f"line {n}: not a JCL statement and not instream data")
            continue
        instream = False
        if len(line) > 71 and line[71:].strip():
            problems.append(f"line {n}: text past column 71")
        body = line[2:71].rstrip()
        if continuing:
            if not body.startswith(" "):
                problems.append(f"line {n}: a continuation leaves column 3 blank")
            continuing = body.strip().endswith(",")
            continue
        statements += 1
        name, rest = ("", body.strip()) if body.startswith(" ") else (body.split(" ", 1) + [""])[:2]
        op, _, operand = rest.strip().partition(" ")
        operand = operand.strip().split(" ")[0]
        if name and len(name) > 8:
            problems.append(f"line {n}: the name {name} has {len(name)} characters: at most 8")
        elif name and not NAME.match(name):
            problems.append(f"line {n}: the name {name} must start with a letter or # @ $")
        if op not in ("JOB", "EXEC", "DD"):
            hint = "did the line before end without a comma?" if "=" in op else "is the name in column 3?"
            problems.append(f"line {n}: {op} is not JOB, EXEC or DD: {hint}")
            continue
        if op == "JOB":
            seen_job = True
        elif not seen_job:
            problems.append(f"line {n}: the deck does not start with a JOB statement")
        if op == "EXEC":
            program = re.search(r"PGM=([A-Z0-9#@$]+)", operand)
            steps.append({"name": name or "(no name)", "pgm": program.group(1) if program else "?", "dds": []})
        elif op == "DD" and name and steps:
            if name in steps[-1]["dds"]:
                problems.append(f"line {n}: step {steps[-1]['name']} has a second DD statement {name}")
            steps[-1]["dds"].append(name)
            instream = operand.startswith("*")
        continuing = operand.endswith(",")
    return problems, steps, statements, data_lines


def assigns(program):
    source = Path(program.lower() + ".cob")
    return ASSIGN.findall(source.read_text()) if source.exists() else None


for deck in sys.argv[1:]:
    problems, steps, statements, data_lines = read(deck)
    print(f"{deck}: {statements} statements, {data_lines} lines of instream data")
    for step in steps:
        print(f"  step {step['name']:8} PGM={step['pgm']:9} DD {' '.join(step['dds'])}")
        names = assigns(step["pgm"])
        if names is None:
            continue
        for missing in [a for a in names if a not in step["dds"]]:
            problems.append(f"step {step['name']}: {step['pgm']} assigns {missing}, and the step has no DD {missing}")
        for unused in [d for d in step["dds"] if d not in names and d not in SYSTEM_DDS]:
            problems.append(f"step {step['name']}: no ASSIGN of {step['pgm']} names the DD {unused}")
    for p in problems:
        print(f"  problem: {p}")
    print(f"  {len(problems)} problem(s)")

daily-posting.jcl: 12 statements, 2 lines of instream data
  step POST     PGM=POSTTXN   DD STEPLIB TXNIN POSTED SYSOUT
  step REPORT   PGM=BRREPORT  DD STEPLIB POSTED BRANCHES BRREPT SYSOUT
  0 problem(s)
posting-bad.jcl: 13 statements, 2 lines of instream data
  step POST     PGM=POSTTXN   DD STEPLIB TXIN POSTED SYSOUT STEPLIB POSTED BRANCHES BRREPORTS SYSOUT
  problem: line 9: DCB=(RECFM=FB,LRECL=22) is not JOB, EXEC or DD: did the line before end without a comma?
  problem: line 11: REPORT is not JOB, EXEC or DD: is the name in column 3?
  problem: line 12: step POST has a second DD statement STEPLIB
  problem: line 13: step POST has a second DD statement POSTED
  problem: line 18: the name BRREPORTS has 9 characters: at most 8
  problem: line 19: step POST has a second DD statement SYSOUT
  problem: step POST: POSTTXN assigns TXNIN, and the step has no DD TXNIN
  problem: step POST: no ASSIGN of POSTTXN names the DD TXIN
  problem: step POST: no ASSIGN of POSTTXN names the DD BRANCHES
  problem: step POST: no ASSIGN of POSTTXN names the DD BRREPORTS
  10 problem(s)

from session import files, run

# The two steps of job DAILYPST run with GnuCOBOL: each DD statement of daily-posting.jcl becomes a DD_ variable that
# names a file of this folder, and each program's exit status is its step's return code.
files("posttxn.cob", "brreport.cob", "txn.txt", "branches.txt")
run("""
cobc -x posttxn.cob
cobc -x brreport.cob
!DD_TXNIN=txn.txt DD_POSTED=posted.txt ./posttxn
cat posted.txt
DD_POSTED=posted.txt DD_BRANCHES=branches.txt DD_BRREPT=brrept.txt ./brreport
cat brrept.txt
# The REPORT step again, with no DD for BRANCHES
!DD_POSTED=posted.txt DD_BRREPT=brrept.txt ./brreport
""")

$ cobc -x posttxn.cob
$ cobc -x brreport.cob
$ DD_TXNIN=txn.txt DD_POSTED=posted.txt ./posttxn
POSTTXN: rejected PUN AC0004 X 000010000
POSTTXN: rejected PUN AC0003 C 00009A000
POSTTXN: read 00006, posted 00004, rejected 00002
[exit status 4]
$ cat posted.txt
MUM AC0001 C 000250000
PUN AC0003 D 000045050
MUM AC0002 C 000120000
MUM AC0001 D 000002500
$ DD_POSTED=posted.txt DD_BRANCHES=branches.txt DD_BRREPT=brrept.txt ./brreport
BRREPORT: 2 branches printed
$ cat brrept.txt
BRANCH       CREDITS      DEBITS
MUMBAI        3,700.00       25.00
PUNE              0.00      450.50
# The REPORT step again, with no DD for BRANCHES
$ DD_POSTED=posted.txt DD_BRREPT=brrept.txt ./brreport
BRREPORT: OPEN statuses 35 00 00
[exit status 12]