Rocket® COBOL

Problem: The 'usercall.c' is no longer in the product. What is its replacement? Resolution: The method of calling C from COBOL using usercall.c has been removed from Server Express product as obsolete. A more effective method is to use the CALL "C_func" functionality, as documented in the chapter 'Mixed-Language Programming' in your Program Development book. Old KB# 7240

Problem: A COBOL program prints 30 tables. Each one is OPEN OUTPUT and under Server Express only the last table is in the print file at the end of execution each individual file has been overlaid by the OPEN OUTPUT command. Resolution: On Server Express it is necessary to code the OPEN statement as OPEN EXTEND. This will cause each additional file to be concatenated onto the end of the file. See the "Language Reference" manual for syntax rules for coding OPEN EXTEND. Old KB# 7225

Problem: A sample clustered system has a file system that's shared between machines. On that shared file system within /usr/opt subdirectory is where the ASLMF license keys were installed.  Both the Server Express Developer product and the Application Server product are installed in the original file system, not on the shared file system. Resolution: Application Server's license databases must reside in /var/mfaslmf and not other file systems or other directory structures like /usr/opt/ . Even in clustered systems the system directory, /var/mfaslmf, will be visible. Old KB# 7212

Problem: In designing a transaction-oriented application that emulates a mainframe legacy system, it would be helpful to be able to restore all COBOL modules to their initial state, as though the entire run-unit had exited and had been re-started, but without the overhead of actually exiting and re-starting.   The function CBL_SUBSYSTEM is supposed to have this capability, but how to call CBL_SUBSYSTEM from C? Resolution: Attached to this Knowledge Base article is a tar file containing an example demonstration.  Included are C and COBOL source files, a script showing how to compile, and a "readme.txt" with information and instructions. Attachments: subsystem_demo.tar Old KB# 7220

Problem: I/O error : file 'AS1S' error code: 9/220 (ANS74), pc=0, call=1, seg=0 220     Attempt to execute more than one SORT or MERGE simultaneously The problem is input procedure is working fine and is releasing all records, but while returning the records it is abending and is throwing the error as mentioned above. What could be the reason for this, since no other sort or merge statements are being executed? 006200 PROCESS-RECORDS. 006400     SORT AS1S-SORT-FILE 006500     ON ASCENDING KEY AS1S-SD-NO 006600     INPUT PROCEDURE IS 3100-INPUT-PROCEDURE 006700     OUTPUT PROCEDURE IS 3200-OUTPUT-PROCEDURE. <snip> 007700 3100-INPUT-PROCEDURE SECTION. 007800* 007900        MOVE 100                 TO AS1S

Problem: Customer has both the development product and Application Server on the same machine, installed in two separate locations (cobol and cobolas)  They have five development licenses and two hundred Application Server licenses installed, but process is running out of available licenses and failing with an A147 system abend. Resolution: Application Server licenses are recorded against a specific installation path.  The mfaslmf userlog showed that the application was trying to use the five development licenses rather than the Application Server licenses.  This was ascertained by examining the error messages in the userlog: No license AS /usr/lib/cobol This showed that the environment was pointing to the development product (cobol), rather than the Application Server installation (cobolas). Firstly check that COBDIR, PATH and LD_LIBRARY_PATH/LIBPATH/SHLIB_PATH are pointing to the correct product location. Check any .profi

Problem: Using Informix File handler for Informix files instead of the Micro Focus file handler. Basically, the number of open files increases to a stage where some processes hangs and when it reaches the open file limit, it crashes.  Common files are not closing, while all the other files are closed, which causes the number of open files to increase. eg: 1. Program 1 opens files A, B, C and D 2. Then Program 1 calls Program 2 which open files X, Y, A and B. 3. Program 2 finishes it process, close files X, Y,  A and B and exit program. 4. Program 1 continues it's process, at this moment Files A and B are opened twice in the system. 5. When this happens several times, the number of files opened by a process exceed the limit and the program crashes with Informix error 24. Resolution: Try the following tunable: multi_close_limit Syntax:                set

Problem: Programs that accept data from command line return different results when run from cobrun or cobjrun. ex. cobrun proga parm1 With cobrun, 'parm1' is returned when doing an ACCEPT ws-var FROM COMMAND-LINE. with cobjrun, 'proga parm1' is returned instead. Resolution: This is the expected behavior...The problem is that when using cobrun, the command line is being passed directly to the COBOL run-time system.  When using cobjrun, the command line is passed to the JAVA Virtual Machine.  In this case, the JVM turns around and creates a new command line to invoke the COBOL run-time system. When using cobrun, argument 0 is the program name itself and your command line parameters start with argument 1 to argument n... When using cobjrun, argument 0 becomes the cobrun command inserted there by the JVM to invoke your COBOL program, argument 1 becomes the program name and your command line parameters start with argumen

Problem: When migrating from Object COBOL Developer Suite (OCDS) to Server Express, linking a Character Dialog program, some objects cannot be found. Like for example: DSC.o DSCOMP.o DSCRUN.o DSCVALRN.o DSDLGINI.o DSERRHAN.o DSFLD.o DSGETSS.o DSLCONV.o DSNLSRTN.o DSRUN.o DSTERM.o DSTRACER.o DSUSRCAL.o DSUXSYSP.o DSVRSCRN.o (These files were in $COBDIR/src/dialog under OCDS). Resolution: On Server Express these files are not needed (they are built into the runtime). All you have to do is to remove all references to those files, and to relink. Old KB# 7218

Problem: What is the guideline of using mfsort or mfsort 64 ?  My application is built as 32Bit mode but it is possible to run mfsort under COBMODE=64.  Does mfsort64 have more capacity or performance than mfsort with COBMODE_32 ? Resolution: There is no difference between mfsort and mfsort64, other than one being 32-bit and the other 64-bit. Old KB# 7244#AcuCobol#netexpress#RMCOBOL#ServerExpress#COBOL

Problem: The client/server binding demo at $COBDIR/demo/csbind is coded so the file that is read by the client is always saved at the server with a fixed file name, "netxcpfl.dat". Because of this limitation, only 1 client can be run at the same time and successfully create the corresponding output file in the server. Resolution: The version of netxcli.cbl and netxserv.cbl attached have been altered slightly so if the user enters  the name of the output file at the client, it is sent to the server and used, otherwise the default name "netxcpfl.dat" is used. Also, when the file has been read completely and the server is still running, there is a new pause in the client. Attachments: netxdemo.zip Old KB# 7251

Problem: The new HOSTARITHMETIC compiler directive available in Server Express can guarantee decimal truncation when ON SIZE ERROR is not specified. Resolution: Does 9 1 = 0?  If a data item can hold only one digit (PIC 9), and currently has the value 9, what should be the result of adding 1? If the resulting value of an arithmetic expression is larger than the maximum value the target item can store, then a "size error" condition exists.  If the ON SIZE ERROR clause is present, then no value is stored in the target, and the imperative statement(s) in the clause are executed. If no ON SIZE ERROR clause is present, then the arithmetic results are undefined.  This rule is documented in the ANSI standard, as well as in Micro Focus and IBM documentation.  Thus, the result of the following code fragment is undefined, and it is a programming error to rely on any specific result: 01 num-ws pic 9 value 9. add

Problem: Problems converting a fixed length line sequential file to variable length idxformat4. The input file is LSEQ 662, fixed length.  The output file is mf4 format variable length 662-4089. Using the command: rebuild infile, outfile -o:LSEQ,IND -t:mf4 -k:1-48 -r:v662-4089 getting the error message 'REBUILD aborted - Invalid key specification - key not contained in record'. The -r option specifies the record format.  However, the input file is fixed 662 and the output file is variable. How do I specify the input and output record formats in the same command? Resolution: Sequential files do not have header information  so the organization and record length must be specified. There is only one keyword for record length (-r) and it is not possible to specify a different record length on the input and output files. To get around this, run the conversion in two steps: Try running the conversion from lseq to

Problem: After migrating an application from VAX to Solaris UNIX and Server Express, one may notice that the last byte in some data items in his output files appeared differently than it did when the program was run on the VAX. Resolution: For numeric data items of usage DISPLAY, the sign is encoded in the last byte (unless the SIGN TRAILING SEPARATE clause is specified for the item).  There are two popular conventions for encoding the sign in the last byte:  the ASCII convention and the EBCDIC convention.  Micro Focus uses the ASCII conventin by default, and apparently VAX COBOL uses the EBCDIC convention. The compiler directive named SIGN can be used to change Server Express's behavior in this regard.  For compatibility with the VAX,  the following option can be specified on the "cob" command used to compile the programs:   -C SIGN=EBCDIC Old KB# 7221

Problem: After installing a new version of our COBOL compiler the breakpoints in Animator between starts of the anim. gets lost Resolution: This is normal behaviour with Server Express as per the documentation, in Server Express an aif file is not created if you use COBSW= A. To preserve the aif file you must use the COBAIFNAME environment variable. export COBAIFNAME=myprog Old KB# 7232

Problem: How to modify the default behaviour for LINE SEQUENTIAL files. Resolution: Various aspects of the File Handler's behaviour are configurable.  You can alter the behaviour by setting various options in the File Handler configuration file (EXTFH).   By Default the File Handler truncates trailing spaces from Line Sequential and Line Advancing Files. The behavior of truncating spaces can be altered by using the following File Handler option. STRIPSPACE When performing WRITE or REWRITE operations on line sequential or line advancing files, the STRIPSPACE option determines whether to remove trailing space characters. STRIPSPACE={ON|OFF} Default:    ON Old KB# 7243#COBOL#RMCOBOL#AcuCobol#ServerExpress#netexpress

Problem: Installing Server Express v2.2, gets message about AIX version saying 'product was built on AIX 4.3.3 but you are running AIX 5.2.0, please confirm you want to continue'. Should he continue?   Resolution: Yes. This is normal and expected in this case, he was installing the 32-bit version of Server Express v2.2. Old KB# 7213

Problem: The behaviour of VARCHAR2 host variables have changed in Pro*COBOL v8.0 You have VARCHAR2 host variables defined as PIC X(n) and are having problems with tests/compares in your WHERE clause after upgrading your version of Oracle. Resolution: Oracle have a configuration option for Pro*COBOL to revert back to the old behaviour. In the $ORACLE_HOME/precomp/admin/pcbcfg.cfg file, add the following: PICX=VARCHAR2 From now on by default, PIC X(n) host variables will no longer have the trailing spaces truncated unless you have this  configuration option. Old KB# 7280

Problem: When using the directive OUTDD(file n) where n is a value less than 95, for example 80, and when using DISPLAY string UPON SYSOUT, the contents of string are split into several lines, where each line length is much less than the value of n and keeps varying (decreasing). Resolution: This issue has been fixed with the latest fixpack for Server Express 4.0 SP1. There is also a workaround: OUTDD($SYSOUT 95 L A) or OUTDD($SYSOUT) - the default values are 132 L A. If you really need the lines split at 94 or less (80 for example), generate the file so no line is split by defining the record length to the maximum value that can be found, and it should be very easy to write a little program that splits them to 80: ...         fd sysoutfile.         01 sysoutline pic x(maxlinelength).         fd sysout80file.   &n

Problem: A situation can exist where it is not possible to create an executable using "cob -x".   The error occurs even when testing one of the simple demonstration programs that are furnished with the product, for example: mkdir  /tmp/test cp  $COBDIR/demo/debug/pi.cbl  /tmp/test cd  /tmp/test cob -x  pi.cbl      The system linker "ld" reports not being able to find a library named "cobstat" or "-lcobstat".   The exact wording of the error message depends on the bit mode (whether 32 or 64), but always mentions "cobstat".   This problem is specific to HP/UX and does not occur on other UNIX variants. Resolution: This problem is caused by the environment variable LDOPTS being set as follows: LDOPTS="-a shared" This causes the system linker to search only for shared libraries (.sl),

Problem: If you have installed Net Express 4.0 or Server Express 4.0 with the Enterprise Server option, then the CCITCP2 program will have been replaced by a trigger program that calls the Micro Focus Directory Server (MFDS) program. As far as CCI users are concerned, the Directory Server serves the same purpose as CCITCP2 and they behave in exactly the same way.  Each machine that runs the Enterprise Server option must have the Directory Server program running on it. MFDS is a complete replacement for CCITCP2, is backwardly compatible with CCITCP2 and supports all the same command line options and syntax and all other functionality. Resolution: There are two types of servers controlled by the Directory Server Micro Focus Enterprise Server This is a server that provides an execution environment for COBOL programs running in response to service requests issued by several different client types. Micro Focus CCI Server This is a server that

Problem: You may encounter a File Status error 9/068 (record locked) on reads of records in an INDEXED file that you know are being read by other processes but you are certain are NOT being LOCKed. For example,  assume several processes are all running the same program. The program has LOCK MODE MANUAL set in the SELECT clause. It opens the file I-O, sets the record pointer using START NOT LESS THAN and then READ's NEXT through the file with out any locking on the reads. Provided there is not a different program accessing that file there should NOT be any record locking take place, but it can. Resolution: Specify the NODETECTLOCK compiler directive. NOTE: This error does not occur if the file is opened INPUT              This did not behave in this manner in Object COBOL Developer Suite               The lock

Problem: Compiler directive HOSTARITHMETIC guarantees decimal truncation on arithmetic overflow when ON SIZE ERROR is not specified.   Does 9 1 = 0?  If a data item can hold only one digit (PIC 9), and currently has the value 9, what should be the result of adding 1? If the resulting value of an arithmetic expression is larger than the maximum value the target item can store, then a "size error" condition exists.  If the ON SIZE ERROR clause is present, then no value is stored in the target, and the imperative statement(s) in the clause are executed. If no ON SIZE ERROR clause is present, then the arithmetic results are undefined.  This rule is documented in the ANSI standard, as well as in Micro Focus and IBM documentation.  Thus, the result of the following code fragment is undefined, and it is a programmatic error to rely on any specific result: 01  num-ws   pic 9 value

Problem: What  is the purpose of .hnf files, which can be found in $COBDIR/demos. Resolution: The files are for use with Hyhelp, the Micro Focus help system. The ones supplied are used within Editor and Animator. Old KB# 7261

Problem: cobscan64 runs the Scan64 utility. The Scan64 utility enables you to check programs for pointer violations. You use Scan64 to help you migrate applications from 32-bit to 64-bit COBOL systems. If you have developed COBOL programs using the Server Express 32-bit development environment that you want to migrate to the Server Express 64-bit development enviroment, you should run Scan64 before compiling and running them with the 64-bit system development environment. Resolution: Note: Scan64 cannot detect invalidly-sized data items provided to prototyped CALL statements that require pointers. You can only use Scan64 on applications that have been compiled to intermediate code. The format of the command is: cobscan64 [-v] [-m] [-ttfile [-ltlist]] filename where the parameters are:   -v         Internal addressing information is displayed. This can be useful to our support  &nbsp