It can only be used in conjunction with an AWM modeled remote edit action (see the tool Edit MVS File and Browse MVS File of the AWM function package). There is no support for COBOL programs opened from MVS Explorer in the Mainframe Systems Explorer view.
The Mainframe Pre-compiler Support is a technique that lets you call language pre-compilers (also known as pre-compilers) on the mainframe while editing a source file on the client machine. The source file is taken from the editor and precompiled on the mainframe. The result is transferred back to the client machine and used as input to the background parsing process, so the editor does not flag syntax errors due to an unknown pre-compiler. All COBOL variables are usable with the Content Assist.
To use the Mainframe Pre-compiler Support, additional configuration is necessary.
A Mainframe Pre-compiler must be configured on the mainframe site to be used with the Mainframe Pre-compiler Support. A pre-compiler specific REXX procedure must be written on the mainframe to connect a pre-compiler. That procedure has to set up the required environment for the pre-compiler. For example, by allocating data sets and then call the pre-compiler.
The pre-compiler exit procedure has to be named TAUZCXPP and must be placed on a library that is in the search path in the user server. TAUZCXPP is called with the following parameters:
| Parameter | Description |
|---|---|
| Indsn | Data set name of a cataloged mainframe file containing the input source for the pre-compiler. This is the content of the editor on the client machine when the background parse process is started. |
| Outdsn | Data set name of a cataloged mainframe file for the output generated by the pre-compiler. This is the translation of the input file as done by the pre-compiler and should be valid COBOL. |
| Msgdsn | Data set name of a cataloged mainframe file for messages generated by the pre-compiler. |
| Chgdsn | Data set name of a cataloged mainframe file for information about changes done by the pre-compiler. |
| Parms | Additional user supplied parameters for the pre-compiler call. |
All data sets exist in the mainframe file system when the procedure is called. The procedure must allocate the input and output data sets to the pre-compiler specific DD names. The message and change data sets are optional and need to have a special format as described in the following sections.
The following return codes of TAUZCXPP are supported:
| Return code | Description |
|---|---|
| 0 | Changes done by the pre-compiler are known and the change data set is filled. |
| 1 | Changes done by the pre-compiler are unknown and TAUZCXPP did not fill the change data set. |
If return code 1 is received, the mainframe pre-compiler interface will try to set up a change data set by comparing the input data with the output data. A simple TAUZCXPP might just generate the output data set and issue return code 1.
Messages from a pre-compiler must be placed in the message data set. The data set is a fixed block data set with a logical record length of 80. It is already cataloged when TAUZCXPP is called.
The local compiler needs to know which lines of the original source code were changed, inserted or deleted by the mainframe pre-compiler. This information may be calculated using the Mainframe Pre-compiler Interface by issuing return code 1 from the exit TAUZCXPP, or may be set by TAUZCXPP itself. The data set is a fixed block data set with a logical record length of 80. It is already cataloged when the exit TAUZCXPP is called.
Changes can be grouped into blocks so that, for example, deleting five consecutive lines is a block of five lines.
There must be one line for each change block:
| Offset | Length | Description |
|---|---|---|
| 0 | 5 | Number of the line in the input source that this change belongs to (right padded filled with 0). |
| 5 | 1 | Reserved (should be blank) |
| 6 | 1 | Type of change |
| 7 | 1 | Reserved (should be blank). |
| 8 | 5 | Length of the change block (right padded filled with 0). |
| 13 | 67 | Reserved (should be blank). |
COBOL compiler options can be set in an AWM model by the AWM application specific attribute COMPOPT.
In order to integrate a mainframe pre-compiler, the COMPOPT attribute must contain the call of the local pre-compiler interface:
p(AWMREMPC) ENDPAfter p(AWMREMPC) all parameters can be passed which are needed by the mainframe pre-compiler. The parameters are passed through to the mainframe pre-compiler exit.
The integration of the mainframe pre-compiler is realized on the client side as a stackable preprocessor implementing the integrated preprocessor interface as described under "Writing a Preprocessor". It is recommended to use the standard CP preprocessor as the first preprocessor in the stack because the CP preprocessor ensures that the contents of unsaved edit buffers is retrieved and passed to the succeeding preprocessors. If it is required that the mainframe pre-compiler integration gets the source from the edit buffers in an unchanged way the CP preprocessor has to be called with the parameters NOSQL NOCOPY. Otherwise the CP preproessor resolves all COPY and EXEC SQL INCLUDE statements in the source before passing the source to the mainframe pre-compiler integration. You find that the COMPOPT attribute typically has the following format:
p(AWMREMPC) p(CP) NOCOPY NOSQL ENDP
You specify all the parameters to be passed to the Mainframe Pre-compiler exit between the p(AWMREMPC) and p(CP).
Another AWM application-specific attribute AUTO_PARSE controls whether the background parse process of the editor is started automatically or only if explicitly called.
Micro Focus recommends that you set AUTO_PARSE to FALSE if a mainframe pre-compiler is configured. In this case, you should model a tool of type “Reparse with Refresh” from the Micro Focus Editor function package, and add it as a related action to the element type representing COBOL programs and as global editor action. This action can be used to trigger the start of the pre-compiler interface from the context menu of the COBOL program and the editor context menu. If the file should be parsed when it is opened in the editor, you should additionally specify the AWM application-specific attribute INITIAL_PARSE and set it to true.
There are two mainframe pre-compiler exit samples provided in the EXEC library of the z/Server delivery:
TAUSADCL is a mainframe pre-compiler exit which resolves the DB2 EXEC SQL INCLUDE statements within a COBOL program. TAUSADCL expects as input the list of libraries separated by semicolon where to look up the SQL includes.
TAUSADCL generates an output file where all DB2 EXEC SQL INCLUDE statements with the exception of SQLCA and SQLDA are commented out and replaced by the content of the include. The TAUSADCL exit produces the change data set on its own. This results in an exit with a return code of 0 if successfully executed. If an SQL include can not be resolved from the provided include libraries TAUSADCL returns a corresponding message in the message data set.
Specify the following format for the COMPOPT option used to call the exit TAUSADCL:
p(AWMREMPC) RESOLVEDCLGEN <list of include libraries> p(CP) NOSQL NOCOPY ENDP ENDP DIALECT(ENTCOBOL) XDB
TAUSADPC is a mainframe pre-compiler exit which calls the IBM DB2 pre-compiler with the source passed by the Mainframe pre-compiler interface. TAUSADPC expects as input the list of libraries separated by semicolon where to look up the SQL includes.
The output file generated by the IBM DB2 pre-compiler is returned to the mainframe pre-compiler interface under the allocation where the pre-compiled source is expected. TAUSADCL exits with the return code 1 on successful execution. This results in the exit leaving it to the mainframe pre-compiler interface to produce the change data set by comparing the input data set with the output data set. There is only a simple error handling implemented in the exit. If the exit is intended to be used in a production environment it has to be extended with adequate error handling.
Specify the following format for the COMPOPT option used to call the exit TAUSADPC:
p(AWMREMPC) DB2PC <list of include libraries> p(CP) NOSQL NOCOPY ENDP ENDP DIALECT(ENTCOBOL)