Treat characters as bytes in STRING, UNSTRING, and INSPECT statements. By default isCOBOL internally converts strings into Unicode. The -b option makes isCOBOL work directly on the string without any conversion. This approach increases performance but programs compiled with this option may not work correctly if they use national items.

The option takes effect on standard alphanumeric data items, while national data items are not affected.

Use the configuration file identified by config_file. See Configuration for the list of the configuration properties that are applicable to the Compiler and for details about how the configuration is built.

Use only the configuration file identified by config_file. See Configuration for the list of the configuration properties that are applicable to the Compiler.

Include debug information.

Enable extended debugger functions. This option implies -d.

In addition to the standard debug features, all the variables in the class are generated, including those not used in the program and the literal constants that are generated during the execution and not as static fields in the generated class. When a program is compiled with -dx, the Debugger is able to query and set all the items of the program Data Division including the items that are not used in the Procedure Division and the IDE allows the source code to be changed while debugging. With -dx the Debugger is able to skip statements through the “jump” command.

Removes output class files if compilation fails. This option works only when the source file has been recognized as a COBOL program or class. If some of the class files cannot be deleted for some reason, the compiler doesn’t signal it. The number of removed classes is shown at the bottom of the compiling result.

Output errors to a '.err' file. The file has the same name as the source and is created only if there are compiler errors or warnings. The Compiler automatically removes the ‘.err’ file before starting to compile the source, so existing err files disappear after a correct compilation. The Compiler output is also traced on the system output.

Specifies the directory for error files. If the directory does not exist or doesn’t have the correct permissions, the err file will not be created.

Stop compilation and return a non-zero exit code if an error occurs. This option is useful when compiling multiple sources at once.

Sets the maximum number of errors printable by the Compiler to n, where n is a positive number.

When multiple source files are compiled at once (e.g. if you use wildcards in the source name), the option limits the number of errors for each single source file, not for the whole compilation.

Display the list of all compiler options with a quick explanation for each one of them and exit.

Display the list of all compiler options including experimental options with a quick explanation for each one of them and exit.

Disable the Automatic compilation of referenced COBOL classes

Optimize EVALUATE with string literals. When a EVALUATE statement tests string literals, isCOBOL uses the Java SWITCH statement instead of the EVALUATE implementation.

Display the Compiler version number and exit.

Display verbose output, e.g. the count of errors, informational and warnings.

Allows you to override COBOL library routines. This option is useful for overriding COBOL library routines such as C$SYSTEM and for speeding up CALLs to subprograms that are called frequently. The -sysc option causes the compiler to place the Java class in the com.iscobol.lib package (or com.iscobol.lib_n package when compiling with -cp). The isCOBOL runtime framework searches for programs in this package before searching other places such as paths specified in iscobol.code_prefix. To execute programs compiled with -sysc, they must be found in a path or jar file listed in the Java class path.

With the -sysc option, the compiler adds "package com.iscobol.lib;" (or "package com.iscobol.lib_n;") to the top of the generated Java source code for the program class. For example, if a program named MYPROG is compiled with -sysc then the generated Java class will be named com.iscobol.lib.MYPROG.

The -sysc option should be used only for CALLed subprograms. If it is used for main programs or COBOL objects there is no error at compile time, but the generated object cannot be executed.

Acucobol compatibility flag:

 

WITH CONVERSION is assumed for MOVEs from alphanumeric items to edited items.

Specifies the default file assignment as external.

Count bytes instead of characters in FIXED/ANSI source files.

This option is useful during the compilation of source files that include double-byte characters. Without this option, text written after column 72 may be considered as written before column 72 due to the double-byte characters in the source line.

Shows USAGE DISPLAY memory content. This option affects the internal definition of variables. If external variables are used, then all programs have to be compiled with this option, otherwise a mismatch error is received. This option shouldn’t be used with programs that display a graphical user interface.

Note that this option affects also the debugging of the program, as the Debugger will show memory content of USAGE DISPLAY items instead of showing their value.

Compatibility setting for file SEQUENTIAL is LINE SEQUENTIAL. When this is not set, files with ORGANIZATION IS SEQUENTIAL are treated as BINARY SEQUENTIAL.

Intermediate results are always calculated to 36 digits.

 

isCOBOL arithmetic uses 3 types of number:

(A) fixed point numbers with number of digits less or equal to 18;

(B) fixed point numbers with number of digits greater than 18;

(C) floating point numbers;

 

Without this option, when an arithmetic operation occurs, the most wide type is used to perform the operation, i.e:

 

(A) +-*/ (A) -> (A)

(A) +-*/ (B) -> (B)

(B) +-*/ (B) -> (B)

(A) +-*/ (C) -> (C)

(B) +-*/ (C) -> (C)

(C) +-*/ (C) -> (C)

 

The option causes that the two operation whose result can be wider than the operands, i.e. division and multiplication, will be performed using the (B) type, i.e.:

 

(A) */ (A) -> (B)

(A) */ (B) -> (B)

(B) */ (B) -> (B)

(A) */ (C) -> (C)

(B) */ (C) -> (C)

(C) */ (C) -> (C)

List keys in offset order. Without this option keys are listed following the order they’re declared in the FILE-CONTROL paragraph.

 

It changes the order in which keys are registered in the physical file, that can be verified with file management utilities such as jutil and ctutil. As a consequence, this option affects XML/ISS dictionaries as well as the I$IO and file interfaces where keys are pointed by ordinal number.

ICOBOL compatibility.

 

Micro Focus compatibility flag. It supports the following:

 

 

In addition, SORT RETURN and SORT MESSAGE internal variables are created. SORT-MESSAGE is never used while SORT-RETURN is checked before every RELEASE and RETURN statement and if it contains 16 the sort is aborted and the control returns to the instruction following the SORT statement.

Microsoft COBOL compatibility flag. It supports the following:

 

 

In addition, LIN and COL internal variables are created. As a consequence the reserved word COL cannot be used.

Use DBCS instead of Unicode in PIC N without USAGE NATIONAL

Leading zeros are shown when numeric data items are displayed on a character based screen.

Changes OCCURS DEPENDING ON behaviour.

This affects data items that appear after a variable-length table in the same record; that is, after an item with an OCCURS DEPENDING clause, but not subordinate to it. With -cod1, these items always immediately follow the table, regardless of its current size; this means their addresses change as the table's size changes.

Closes all files opened by the program when the program exits.

 

This option has priority over the iscobol.file.close_on_exit (boolean) configuration setting. For example, if a program is compiled with -coe, files will be closed even if it sets iscobol.file.close_on_exit=false with a SET ENVIRONMENT statement before exiting.

Enable full pointer support.

 

Use this option if you need to call C functions that reuse pointers.

 

Programs compiled with this option can run along with programs compiled without this option in the same COBOL application, but they cannot share parameters each other. Configuration properties can be used to share information between these two different kind of programs.

 

Programs compiled with this option can’t call C functions with CALL CLIENT statements in a thin client environment.

 

If programs try to allocate or reference items out of their bounds, consider using -m1 as well, otherwise the whole JVM may crash.

 

Pointers are 4 bytes in size by default. Use -d64 to make pointers 8 bytes in size.

Allow ++INCLUDE statements. These statements are internally translated to COPY statements.

RM/COBOL compatibility flag. It supports the following:

 

RM style lock mode: any READ LOCK on a file that does not have an applicable DECLARATIVE section is automatically translated into READ LOCK WAIT.

List keys in offset order in RM/COBOL compatibility. Without this option keys are listed following the order they’re declared in the FILE-CONTROL paragraph.

 

This option differs from -cko on split keys composed of multiple segments. When using -cko, only the first segment of the key is evaluated for the offset. When using -crko, instead, all the segments of the key are evaluated for the offset.

 

This option changes the order in which keys are registered in the physical file, that can be verified with file management utilities such as jutil and ctutil. As a consequence, this option affects XML/ISS dictionaries as well as the I$IO and file interfaces where keys are pointed by ordinal number.

Compatibility setting for implicit record varying size for files with multiple record definitions with different lengths and files whose record is varying from size to size. When this is not set, files are treated as fixed length and the maximum record length is used.

 

This option affects files that include two or more record definitions with different size as well as files that include OCCURS DEPENDING.

Treat the COBOL name in ASSIGN clause as a literal. This allows a mapping to be created for the file name if the iscobol.file.env_naming (boolean) configuration property is set to true.

Compatibility setting: SQL returns an error if a host variable is set to null.

 

The error number stored in SQLCODE is 1405. It can be customized by setting iscobol.esql.value_sqlcode_on_null, but not by setting iscobol.esql.sqlcode.1405.

Quotes inside ESQL statements are left as they are by the Compiler. Without this option, all kinds of quotes are translated to single quotes by the Compiler.

Treats numeric USAGE DISPLAY data as characters in comparisons and moves.

 

This option affects the comparison between numbers whose usage is DISPLAY in particular cases, using a byte by byte comparison instead of comparing the numeric representation.

 

The byte by byte comparison is used when:

 

 

This option affects the MOVE statement when one operand is an unsigned numeric data item USAGE DISPLAY, a numeric constant or a numeric literal and the other one is an unedited alphanumeric item.

 

IBM COBOL compatibility flag. It supports following syntaxes:

 

 

 

 

 

 

 

IBM arithmetic compatibility.

 

If the dest-item of a calculation does not include the decimal part, the result of internal operations made to set that result lose their decimal part as well.

For example:

compute res = (11/4) * 4

If res is declared as PIC 99, it will be set to 8

If res is declared as PIC 99v99, it will be set to 11.00

If the -cva option is not used, the result of the above calculation will always be 11.

IBM arithmetic compatibility with powers treated differently.

 

If the expression of the COMPUTE statement contains a power whose exponent is:

 

then standard -cva rules are not applied.

For all other cases, the behavior is the same as when the -cva option is set.

 

Binary data whose length is <= 2 are stored in 1 byte

Treat BINARY as COMP-5.

 

Note that only items explicitly defined as BINARY are affected. COMP and COMP-4 are not affected despite they’re equivalent to BINARY. In order to treat COMP and COMP-4 as COMP-5, use -rm=newmeaning,word..., e.g. -rm=COMP-5,COMP-4.

Use 64-bit pointers for USAGE POINTER data items. This option should be used only in conjunction with -cp.

 

Without this option, pointers are 4 bytes in size.

With this option, pointers are 8 bytes in size.

Use ACUCOBOL numeric format. The compiler uses this convention by default if no other convention is specified.

See USAGE clause for details about how numeric data items are affected by this option.

Use MBP COBOL numeric format

See USAGE clause for details about how numeric data items are affected by this option.

Use Data General numeric format for binary items.

See USAGE clause for details about how numeric data items are affected by this option.

Store any data item whose underlying type is binary in the minimum number of bytes needed to hold it. Normally, binary types are stored in two, four, or eight bytes. This option can be used to emulate the ACUCOBOL -dm option.

Use IBM sign encoding and IBM COMP sizes. COMP sizes are 1 (only if -d1 option is also used), 2, 4, 8, 12 or 16 depending on the item picture.

See USAGE clause for details about how numeric data items are affected by this option.

Use IBM sign encoding and IBM COMP sizes. COMP sizes are 2, 4, 8 or 16 depending on the item picture.

See USAGE clause for details about how numeric data items are affected by this option.

Use Micro Focus sign encoding and Micro Focus COMP sizes.

See USAGE clause for details about how numeric data items are affected by this option.

Use Micro Focus sign encoding and IBM COMP sizes (like MF -C IBMCOMP). This option has the same effect of -dcm except that the length of COMP items is calculated in the same way as -dca. See USAGE clause for details about how numeric data items are affected by this option.

 

In addition, the SYNCHRONIZED clause affects also group items.

Use NCR COBOL numeric format.

See USAGE clause for details about how numeric data items are affected by this option.

Use Realia sign storage convention.

Sign information for S9(n) variables is stored using the conventions for Realia COBOL, and their conversion to binary decimal is the same as that performed by the Realia compiler.

Initialize values of WORKING-STORAGE SECTION data items and indexes by type.

 

 

This option does not affect items declared with VALUE or EXTERNAL clauses or those subordinate to a REDEFINES phrase.

Initialize values of WORKING-STORAGE SECTION data items and indexes by type.

 

 

This option does not affect items declared with VALUE or EXTERNAL clauses or those subordinate to a REDEFINES phrase.

USAGE DISPLAY numeric items with no SIGN clause are treated as if they were described with the SIGN IS TRAILING SEPARATE clause.

Initialize each otherwise undefined byte in WORKING-STORAGE SECTION and FILE SECTION to the specified value when a program is first loaded or canceled and then called. char is the decimal representation of the character. For example, to fill the item memory area with 'A' use -dv=65. Use -dv=0 for low-values and -dv=32 for ASCII spaces.

 

Note that the -dv option does not affect data items declared with VALUE or EXTERNAL clauses. In order to initialize EXTERNAL data items, use either -dvext or -dvexta.

 

If -dv is omitted then the compiler behaves as if -dv=32 was specified (i.e. data items specified without a VALUE clause are filled with ASCII spaces by default).

 

Note that when compiling with -di the value specified with -dv affects only alphabetic, alphanumeric, alphanumeric edited and numeric edited items including those that are declared FILLER. When used with -di, -dv does not affect numeric, pointer or index items. See the -di option for more information.

Initialize each otherwise undefined byte of EXTERNAL data items in WORKING-STORAGE SECTION and FILE SECTION to the specified value when a program is first loaded or canceled and then called. char is the decimal representation of the character. For example, to fill the item memory area with 'A' use -dvext=65. Use -dvext=0 for low-values and -dvext=32 for ASCII spaces.

 

If -dvext and -dvexta are omitted then the compiler behaves as if -dvext=0 was specified (i.e. EXTERNAL data items specified without a VALUE clause are filled with low-values by default).

Initialize each otherwise undefined byte of EXTERNAL data items in WORKING-STORAGE SECTION to the specified value when a program is first loaded or canceled and then called. char is the decimal representation of the character. For example, to fill the item memory area with 'A' use -dvexta=65. Use -dvexta=0 for low-values and -dvexta=32 for ASCII spaces.

 

If -dvext and -dvexta are omitted then the compiler behaves as if -dvext=0 was specified (i.e. EXTERNAL data items specified without a VALUE clause are filled with low-values by default).

Relax size-checking rules.

 

When this option is in effect, the values that can be held in binary and packed-decimal data types are limited only by the number of bytes of storage. The picture is not used for determining the largest value that these types can hold, and when moving to a nonnumeric destination the largest possible value determines the number of digits moved.

Relax size-checking rules in compatibility with Micro Focus NOTRUNC directive.

 

When this option is in effect, the values that can be held in binary data types are limited only by the number of bytes of storage. However, the PICTURE is used when moving data from a binary number to a nonnumeric data item.

Relax size-checking rules in compatibility with Micro Focus TRUNC”ANSI” directive.

 

Each numeric data item stores values up to its PICTURE in size, but COMP-5 items ignore the PICTURE when determining the largest value they can hold. However, COMP-5 items do use their PICTURE when moving a value to a nonnumeric data item.

Create the External File Description XML file(s) for all the files described in the program.

Create the External File Description ISS file(s) for the indexed files described in the program.

ISS files are required by some ctutil functions when working with c-tree. The iscobol.sqlserver.iss (boolean) feature needs these files as well.

 

See External File Description dictionaries for more information.

Create the External File Description XML file(s) for the indexed files described in the program.

 

See External File Description dictionaries for more information.

Specifies the directory for EFD files. If the directory does not exist or doesn’t have the correct permissions, the EFD file will not be generated.

Single record locking is default for files WITH ROLLBACK.

 

Normally, WITH ROLLBACK causes multiple locking rules to be in effect for a file. When this option is used, the WITH ROLLBACK clause does not affect whether single or multiple record locking rules are followed. Single locking becomes the default. You may enable multiple locking either by specifying WITH LOCK ON MULTIPLE RECORDS in a file's SELECT statement or by using APPLY LOCK-HOLDING ON file in the I-O CONTROL paragraph.

Specifies a Unicode-enabled sequential access mode file handler for LINE SEQUENTIAL files. With this option, text files are read and written using the Java classes java.io.FileReader and java.io.FileWriter which access files sequentially rather than in random access mode and also preserve Unicode characters. This option should be used when reading or writing device files and pipes (i.e. files that are not disk files). This is to avoid illegal operations and to properly convert between the Java internal format (i.e. Unicode) and the desired external format. Note that the external format can be specified using the Java file.encoding property.

 

The -flsu option is also useful when programs share sequential files between platforms with different line separators (e.g. program A creates the file on Linux and program B must be able to read the file on Windows).

 

The -flsu option causes sequential files to be assigned to PRINT if no other assignment is specified.

 

When using this option, OPEN I-O, REWRITE and READ PREVIOUS are not supported for LINE SEQUENTIAL files.

 

The encoding is controlled by the Java file.encoding property and not by iscobol.encoding *

LOCK MODE IS MANUAL is implied.

This option has priority over -cm and -ci in terms of default lock mode.

All RECORD SEQUENTIAL files have variable-length records. The Compiler assumes that the FD includes RECORD CONTAINS 1 TO n CHARACTERS clause, where n is the length of the largest record description in the FD. Explicit RECORD or VARYING clauses in the FD are ignored.

Generate the ‘.class’ file. This is the default behavior unless -jj option is used.

Generate the ‘.java‘ file. By default the .java file is removed after a correct compilation. Use this option to keep the .java file on disc. If this option is not used in conjunction with -jc, the Compiler will generate only the .java file and not the .class file.

Passes the specified options to the 'javac' compiler. Multiple values must be separated by spaces.

 

Example:

 

See also iscobol.compiler.javac.options

Use this option along with -lf in order to output full listing to a '.list' file in ANSI format. The list file contains all the source code, all the copybooks are merged into it (unless the SUPPRESS clause is used in the COPY statement) and in most of the cases it can be compiled as it is a standard COBOL program.

If used along with -ld, only the source part is generated in ANSI format, the datamap is always in FREE format.

 

This option is guaranteed to work correctly only if the original program is already in ANSI fixed format, otherwise results are unpredictable.

 

Note: The listing is generated before the syntax analysis and every dot out of quotes is considered as end of the statement, therefore, if you don’t enclose copybook names between quotes, you might obtain an uncompilable list file.

Output full listing and data map to a '.list' file. The list file contains all the source code, all the copybooks are merged into it (unless the SUPPRESS clause is used in the COPY statement). The datamap information is stored at the bottom of the list file and provides the following information for each data item described in the program Data Division: source line, item name, offset (in case the item is part of a group item), physical length, section in which the item is defined, type flags, item type and how the item is referenced in the Procedure Division.

 

The datamap is not generated for CLASS-ID programs.

 

Note: The listing is generated before the syntax analysis and every dot out of quotes is considered as end of the statement, therefore, if you don’t enclose copybook names between quotes, you might obtain an uncompilable list file.

Output full listing to a '.list' file. The list file contains all the source code, all the copybooks are merged into it (unless the SUPPRESS clause is used in the COPY statement) and in most of the cases it can be compiled as it is a standard COBOL program written in free format.

 

Note: The listing is generated before the syntax analysis and every dot out of quotes is considered as end of the statement, therefore, if you don’t enclose copybook names between quotes, you might obtain an uncompilable list file.

Creates only a full listing of the program. This option is the same as -lf except that the Compiler doesn’t compile to a Java class, it just generates the listing file and exits.

Specify the directory where ‘.list’ files are to be stored. If the directory does not exist or doesn’t have the proper permissions, the list file is not generated. This option forces the generation of list files even if -lf was not used.

Put all of WORKING-STORAGE into a contiguous block of memory.

Specify the output directory for classes. If the directory does not exist or doesn’t have the proper permissions, the compilation will fail.

Discard variable names from object files. Variable names are stripped from the compiled object. This option helps save memory and sometimes increases performance, however, exception messages shown by the JVM are less clear and the program cannot be compiled in debug mode.

Non-recursive PERFORM, RM/COBOL style

Recursive PERFORM, Micro Focus COBOL and ACUCOBOL-GT style (default)

Pseudo non-recursive PERFORM, OS/VS COBOL style

Change reserved words. Multiple values must be separated by commas.

Single words (such as DISPLAY, ACCEPT, or ADD) can be changed, but complex statements (such as READ PREVIOUS or NEXT SENTENCE) cannot. For example, -rc=ACCEPT,GETDATA treats the word GETDATA as ACCEPT.

 

It’s not possible to replace a keyword with another keyword using this option. Use -rm if you need to replace a keyword with another keyword.

Change the meaning of reserved words. Multiple values must be separated by commas. This option allows you to deem NULL as LOW-VALUES, COMP-5 as COMP, etcetera. For example, rm=LOW-VALUES,NULL treats NULL as LOW-VALUES.

Suppress reserved words. Multiple values must be separated by commas. This is useful when one or more keywords are used as item names. For example, if the program contains the following variable: 77 PRINTER PIC X(32). , the following option is necessary in order to compile it correctly: -rw=PRINTER.

Causes figurative constant QUOTE/QUOTES to be evaluated single quotes.

Use this option to compile big programs.

 

Several classes are generated. The number of generated classes is conditioned by the iscobol.compiler.max_constants * and iscobol.compiler.max_paragraphs * configuration properties.

 

This option is useful to avoid the "too many constants" error that may appear when compiling huge programs.

In rare cases, this option can be used in conjuncion with -sns=Statements in order to get rid of the "code too large" error.

 

Programs compiled with this option are not optimized, so it’s suggested to use this option only for those programs that actually require it. You can activate the option using the IMP OPTION Directive in the first line of the source file in order to avoid a dedicated compile command.

Set the default extension for source and copybooks. When it is specified, any source or copy library file name that does not explicitly specify an extension has the default extension appended to it. Multiple values must be separated by the appropriate system path separator (such as ";" for Windows, or ":" for UNIX).

Allow the compilation of the specified EXEC macro.

 

-exec=html enables the HTML compiler with the following limitations:

Skip EXEC statements.

Level 78 implies the end of the previous 01 group item.

Force Fixed (aka ANSI) source format.

 

The same effect can be obtained via the SOURCE Directive, specifying >>SOURCE FORMAT FIXED at the top of the source file.

Force all CALL statements to be static. In order to correctly compile a program with this option, all programs called by the program you’re compiling must be available in the CLASSPATH. During the runtime session CALLs will perform better, but classes will always be loaded from the CLASSPATH and never from iscobol.code_prefix paths.

Converts copyfile names to lower case.

Converts copyfile names to upper case.

Allows the currency sign to be changed at runtime. It works in conjunction with the iscobol.runtime.currency * configuration property.

Allows the DECIMAL-POINT clause to be reverted at runtime.

When a program is compiled using this option, the absence or presence of the clause DECIMAL POINT IS COMMA is used only to retrieve the position of the thousands separators and decimal separator inside edited pictures and numeric literals.

The actual character used in any display or print statement at runtime is controlled by the configuration property

iscobol.runtime.decimal_point_is_comma (boolean) * .

Supports the syntax:

and

Environment variables in COPY file names are resolved.

Forces Free source format.

 

The same effect can be obtained via the SOURCE Directive, specifying >>SOURCE FORMAT FREE at the top of the source file.

Allows AREA B to extend to the end of the line, regardless of line length, in Fixed (aka ANSI) format.

 

The same effect can be obtained via the IMP MARGIN-R IS AFTER END OF RECORD Directive, specifying >>IMP MARGIN-R IS AFTER END OF RECORD at the top of the source file.

Allows the mixing of source files and copybooks written in Fixed and Terminal formats. This is useful when you’re writing the main source code using the Fixed format and you need to insert some copybooks that are written in Terminal format, or vice versa.

 

The rules that the compiler uses to determine the source format are as follows:

 

The compiler looks at the character in the 7th column of the first not empty line. If this character is a valid indicator (*,$,/) or a blank, then the source file is assumed to be in Fixed (aka ANSI) format, otherwise it is assumed to be in Free format. If the first character of the first not empty line is a valid indicator, not including the case of "*>", then the source file is assumed to be in Terminal format.

As result, a valid comment on the first line establishes the format in a portable way.

 

The -smat option has two effects:

 

1) If the file is not in Fixed (aka ANSI) format then the file is in Terminal format

2) The analysis is repeated again for each copy file

Use this option to avoid "code too large" compiler errors or to optimize the compiled class file. This option sets the maximum number of COBOL statements that the compiler should attempt to keep together in a single generated Java method. A reasonable value for this option is 200. Try lower values if the "code too large" error persists.

 

In rare cases, the -big option has to be used as well in order to get rid of the "code too large" error.

 

Reason for the "code too large" error: one of the static constraints on the instructions in Java virtual machine code in a class file is that the bytecode size of a single method must be less than 65536 bytes. For COBOL this means that the compiler code generator must split large COBOL paragraphs and sentences into multiple Java methods. See https://docs.oracle.com/javase/specs/jvms/se8/html/jvms-4.html for more information.

Specify all paths in which COPY files can be found. Multiple values must be separated by the appropriate system path separator (such as ";" for Windows, or ":" for UNIX).

 

The location of isCOBOL’s default copybooks can be specified with this option. If not specified, the Compiler automatically searches for them in the directory "../sample/isdef", resolved from the location of the iscobol.jar library. For more information about isCOBOL’s default copybooks see Copybooks.

Converts subroutine names to lower case.

Converts subroutine names to upper case.

Forces Terminal source format.

 

The same effect can be obtained via the SOURCE Directive, specifying >>SOURCE FORMAT TERMINAL at the top of the source file.

Set the length of a tab character (the default value is 8). Multiple lengths are allowed, for example: “-stl=4,8” sets the first tab to 4 characters in length, and the rest to 8 characters in length.

Forces Variable source format.

 

The same effect can be obtained via the SOURCE Directive, specifying >>SOURCE FORMAT VARIABLE at the top of the source file.

AUTO assumed on all ACCEPT statements.

Treat simple ACCEPT and DISPLAY statements in accordance with ANSI semantics. Specifying this option is the same as specifying FROM CONSOLE for all simple ACCEPT statements and UPON CONSOLE for all simple DISPLAY statements. You can change this behavior for individual ACCEPT or DISPLAY statements by specifying an explicit FROM/UPON phrase.

HIGHLIGHT assumed on all ACCEPT and DISPLAY statements.

WITH UPDATE assumed on all ACCEPT statements.

Allows exception keys to be entered by the user for any ACCEPT statement. This option assumes ON EXCEPTION CONTINUE for each ACCEPT that does not specify the ON EXCEPTION clause.

Show warnings for MOVEs of alphanumeric items to numeric items.

Show warnings for features that are currently not supported by WebDirect. This option helps the programmer understand how their program will behave when running with WebDirect.

Show warnings for possible divide by zero without ON SIZE ERROR.

Show warnings for statements that are not supported by EIS/Mobile.

Show warnings for LINKAGE/USING mismatch, if there are parameters that are defined in the program LINKAGE SECTION but not in the PROCEDURE DIVISION USING phrase.

Show warnings for long variables in MOVE WITH CONVERSION.

 

It affects also the MOVE with conversion from alphanumeric to numeric edited, hence warnings may be shown also for normal MOVE from alphanumeric to numeric edited if -caec is used in addition to this option.

Extends the REDEFINES TOO LONG warning also to group items.

Without this option the warning is returned only when both the redefined item and the redefining item have a picture.

Show warnings for reference modifiers out of range.

 

Note: with this option the error #173 Reference modifier out of range is returned as a Warning instead of a Severe error.

Show warnings for variables that are not used. This option is useful for programmers that wish to perform code cleaning by removing useless variable definitions. Consider that by default, the compiler does not allocate data division items that are not used, so this option is useful only to reduce the number of source lines and not the memory usage.

Prints tasks in the compiler output.

All comments starting with “todo” are considered tasks.

Automatically execute the program when the compilation is finished.

Set the program to INITIAL. RESIDENT programs are not affected by this option.

Use 4-digit year in ACCEPT FROM DAY/DATE. Treat ACCEPT FROM DATE as ACCEPT FROM CENTURY-DATE and ACCEPT FROM DAY as ACCEPT FROM CENTURY-DAY