Compiler Directives / Options

Data type compatibility with various environments

(n: 0=EC, 2=Acu, 4=MPE, 5=MF, 6=RM, 7=IBMA, 8=IBME)

IBMA = IBM ASCII (e.g., IBM Power Series, AIX)

IBME = IBM EBCDIC (e.g., IBM Z, z/OS)

-dt:ibm

-dt:ibme

-dt:quote

-dt:q

Compile COBOL to specified Java version and use javac -source 1.n

and -target 1.n options (also see -script:javac option)

Use parameter as generate Java package name causing Java output to

contain the package statement and be placed in a corresponding

directory hierarchy

Generate a Java import statement with the given parameter class specification

(e.g., javax.servlet.*).

-out:indexcheck

-out:ssr

-out:noindexcheck

-out:nossr

Generate OCCURS arrays exactly the size specifies in the DEPENDING ON variable

within enclosing groups

Moves between Groups containing OCCUR DEPENDING ON VAR use the source VAR size.

Note that -dt:ibma and -dt:odoslide should be used for this directive to be effective

Generate shorter Java names for groups and elements, do not generate shared

string constants, implies -out:cobolinjava

Generate an "ambiguous identifier" error when duplicate Level 88 condition names

are referenced (default)

-run no_pageline=ddname1,ddname2,.......

-run no_pageline

Specified files(ddnames) will not increment a line when a page advance is invoked on the file

Not specifying ddnames will imply the same for all output files

ACCEPT and DISPLAY statements (without an explicit UPON clause) default to CONSOLE

( graphical user interface window or ASCII terminal) rather than SYSIN/SYSOUT

(standard in / standard out).

The default is -run:system

ACCEPT and DISPLAY statements (without an explicit UPON clause) default

to SYSIN/SYSOUT (standard in / standard out) rather than
CONSOLE (graphical user interface window or ASCII terminal).

Produce tracing messages for PARAGRAPH/SECTION, CALL, GOTO,

PERFORM, FILE I/O operations & SQL I/O operations

Generate code that sets the run-time option name to value as if -Dname=value

was specified on the Java command line, as an environment variable,

or in the cblconfig or etp_deploy_settings.properties

run-time configurations files.

Adds a run-time parameter for the EBCDIC codepage that will be used if the 'collateAsEBCDIC' flag (can be seen below) is up. 

<number> can be either 037 or 277. 

These numbers match exactly collation sequences:

• SQL_EBCDIC037_CP1_CS_AS
• SQL_EBCDIC277_CP1_CS_AS

If a different codepage number is provided, 037 will be used as default.

Note that there is a slight difference between  Codepage037  and SQL_EBCDIC037_CP1_CS_AS.

This is true for codepage 277 as well.

-cache:auto | enable | disable

| forbid | smart | external

-cache:0 | 1 | 2 | 3 | 4 | 5

-cache:external:section-type

Enable or disable various forms of group and element value caching.

Disable cache is the default which means element values are fetched

from the COBOL section memory (e.g., working storage, linkage section, etc.)

when referenced and saved when changed. Other settings will cache

those values in Java objects and invalidate/flush the values when necessary.

Based on the usage and frequency of REDEFINES, OCCURS,

OCCURS DEPENDING ON and access/modification of groups and

elements within groups you may find that different caching algorithms

(enable 1, forbid 3, smart 4) may perform better in certain instances.

Using external 5 will cause memory allocation to be performed outside

the Java heap. This has two effects, larger programs may benefit by

having their working-storage and other sections allocated in the non-JVM

garbage-collected heap usually reserved for OS malloc() and free()

memory allocation and, when used with UNSAFE calling convention,

it allows for persistent (non changing) pointers to COBOL variables to be

passed to C-libraries invoked from COBOL. Setting auto attempts to

determine the best for your program. Indicating a :section-name

will enable that cache setting only for that section-type (wrk, lnk, fil, etc.).

SQL workarounds for Microsoft SQL Server, specifically define cursors as scroll type SENSITIVE

if they are also UPDATABLE because MS SQL Server disallows the combination

of INSENSITIVE and UPDATABLE.

Use JDBC CONCUR_READ_ONLY for the concurrency setting when defining cursors

(default UPDATABLE cursors)

Schema of the stored procedure. Used at compile time stored procedure

validation against the database. This parameter is optional, as

the -sql:user will be used as the default.

Name of the stored procedure. Used at compile time stored procedure

validation against the database. This parameter is required

for COBOL stored procedure compiles.

SQL Process statements and return results in given parameter mode

(ANSI, ORACLE, DB2). DB2 mode simulates DB2 for z/OS on other

databases such that error DB2 SQLCODEs are translated from

SQLSTATEs and SAVEPOINT-protected INSERT/UPDATE/DELETE

statements do not abort transactions upon error

-sql:optindextimestamp <param>

-sql:protect_include <param>

Protect SQL with an autosave and rollback in case of exception. Param values can be 

• ALL
• Specific types of SQL Query such as SELECT|DELETE|UPDATE
• Identifiers for the SQL statement in a program which can take the following forms:

ProgramName,SectionName,ParaName,SQL statement sequence #

SectionName,ParaName,SQL statement sequence #

ParaName,SQL statement sequence #

For e.g.: 

-sql:protect_include ALL

-sql:protect_include UPDATE

-sql:protect_include DROP

-sql:protect_include program1,sectionA,paraB,4

-sql:protect_include paraC,4

-sql:protect_exclude <param>

Protect SQL with an autosave and rollback in case of exception except for the indicated params

Its supports the following params:

• Specific types of SQL Query such as SELECT|DELETE|UPDATE
• Identifiers for the SQL statement in a program similar to -sql:protect_include as shown above
• SQLCODES n1,n2,n3,.....

For e.g.: 

-sql:protect_exclude SQLCODES  -204,305

Do not allow rollback for sqlcodes -204 and 305

e.g:

-sql:protect_include ALL

-sql:protect_exclude SQLCODES -305,456

 The above two directives will ensure rollback for all sql exceptions except for -304 and 456

Note that one can specify the directive at the environment level by passing the following vm args 
-Dsqlcode_norollback=n1,n2,n3.....
However the compiler directive has precedence and will overwrite any SQLCODES passed thru VM args

-sql:statemapperclass <param>

The parameter is a classname of a class that implements the mapping of SQLCODE/SQLSTATE and Messages between DB.

e.g. Oracle and DB2 when migrating from DB2 though the Oracle code/states/messages need to be DB2

This can be a sparse mapping such that only specific mappings need be specified

com.heirloomcomputing.ecs.exec.SqlStatesFileMapper is an implementation of this class that supports the mappings via a TSV input file

com.heirloomcomputing.ecs.exec.SqlStatesFileMapperV2 is a further implementation of this class that providing additional mappings based on the sql verb executed (e.g. DB2 has returns different codes for certain UPDATE/INSERT and DELETE where as Postgres returns just one).

-sql:hourglass

 -sql:identityinfo

Specify Tables and corresponding Identity Columns

For e.g.:

 -sql:identityinfo tableA:colA

 -sql:identityinfo tableB:colB

 -sql:identityinfo tableA:colA,tableB:colB,tableC:colC

-sql:dateformat <param>

where param is of the various date formats such as 

dd/mm/yyyy

Used to specify date format for IBM DB2 CHAR function during compilation

-sql:returncodezero

Setting this flag will cause the compiler to generate code that will

set the COBOL RETURN-CODE to 0 after a non error SQL statement execution.

Implies -opt:large. Break up output Java classes every COBOL paragraph.

Output classes are also named after the paragraph name.

Implies -opt:large. Break up output Java classes  every COBOL procedure division section.

Output classes  are also named after the section name.

(Default) Generates a optimized version of the initialize statement.

A single java initialize statement is created corresponding to

each COBOL initialize statement for group fields

Generates a non optimized version of the initialize statement.

Multiple java initialize statements is created corresponding to

each COBOL initialize statement for group fields

-opt:staticconstantcount

Fine tune the number of static initializer per java class.  The default is 5000.

This option allows adjustment of that number. 

Alternatively -opt:large will provide a fix that will scale with program growth. 

Static initialize java elements are used for constants,

level 88s and other constant data types

Optimize OCCURS arrays (tables) such that no Java array is actually

generated, but simulated  as an object with getters and setters.

This is faster for sparse arrays, i.e., large OCCURS x TIMES where x

is large but only a few of the elements are actually used.

Also required when redefining a table and subsequent indexes are out-of-bounds.

By default COBOL programs use the ASCII collating sequence to do string comparisons,

in this mode "1" < "a". This flag sets the compiled program to use the

EBCDIC collating sequence which means that "1" > "a". 

Alternatively it is possible to use the COBOL configuration section to set this

collating sequence, this also gives more options to provide custom collating sequences.

configuration section.
object-computer.
heirloom-pc
program collating sequence is mainframe.
special-names.
alphabet mainframe is ebcdic.