Rocket® COBOL

This article explains how to force the unixODBC driver to write a trace file when a program uses AcuSQL runtime to access a database. Problem: To use an ODBC-compliant database (such as MySQL) on Linux, we need to configure unixODBC on the server. If you run into problems doing this, it is possible to create a unixODBC trace file by setting up odbcinst.ini in a manner like this: mysql@linux:/home/mysql$ more /usr/local/etc/odbcinst.iniySQL ODBC 3.51 Driver]Driver = /opt/mysql-connector-odbc-3.51.25r1093/driver/.libs/libmyodbc3.soSetup = /opt/mysql-connector-odbc-3.51.25-linux-x86-32bit/lib/libmyodbc3S.soUsageCount = 1 DBC]TraceFile = /tmp/odbc.traceTrace = yesUsageCount = 1 This generates an ODBC trace file and writes in it each ODBC operation performed with tools like "isql." There is a problem with this, however; operations performed through AcuSQL runtime do not generate and access this trace file.

This article describes how to translate French Characters between a Windows UI screen Character set and the Oracle Database Character Set. Problem: Using a graphical interface on Windows, the data you recover is not stored properly on your Oracle database using ProCOBOL, and you are losing special characters like "éàù". Resolution: Set up the following environmental variable before executing the runtime: Export NLS_LANG=FRENCH_FRANCE.WE8MSWIN1252 This tells the PROCOBOL code you have in your application that the IO with the COBOL program will be made with this Character set and that the data should be transformed before being stored on the database (to whichever Character set the database is set at). Your UI will now handle the standard Windows French characters properly and translate them when communicating with the Database. Old KB# 14259

Do the extend products work on multi-CPU machines taking advantage of the multiple CPUs or multiple cores in a CPU? Problem: Do the extend products work on multiple CPU machines? Or take advantage of multi-core CPUs? Resolution: The extend products do not take advantage of multiple CPUs or multi-core CPUs. Any processor or other load balancing assignment is handled by the Operating System on which the extend products are being executed. There are no plans to enhance the extend product line to implement this type of functionality. There have been no issues reported using the extend products on multiple CPU or multi-core CPU machine that are specific to operating in those environments. Old KB# 14231

Runtime fails to initialize properly. Problem: When trying to execute the 64-bit Windows runtime you would see the following error: The application failed to initialize properly (0xc0150002) Resolution: This is because atermmgr.dll is looking for a 32-bit version of comctl32.dll instead of the 64-bit version. You will need to extract the files in the attached zip archive and replace the existing files in the \\AcuGT\\bin directory Incident Number: 2259559 Old KB# 14204

This KB article demonstrates how to replace trailing spaces in an alphanumeric variable with low-values or other characters. Problem: You have an alphanumeric variable with trailing spaces and you want to change them to low-values or to another character. Resolution: Simply use the following syntax from the INSPECT verb: INSPECT WS-ITEM REPLACING TRAILING SPACES BY LOW-VALUES. Old KB# 14260

This article discusses a case where using the MULTIPLE verb and the binary math package could result in incorrect values. Problem: If a COBOL program used the GIVING clause of the MULTIPLY statement and the destination field contained more post-decimal digits than any of the source fields, it was possible for the compiler to generate code that resulted in incorrect values being returned by the multiplication operation when using the binary math package. For example: WORKING-STORAGE SECTION.01 WS-NUMERIC-FIELDS. 03 A PIC SV9(3). 03 B PIC S9(3). 03 C PIC S9(3)V9(11). PROCEDURE DIVISION.MOVE -0.123 TO A.MOVE 12 TO B.MOVE 12 TO C. MULTIPLY A BY B GIVING C. In this example, fields A and B are the source fields and contain three post-decimal digits each. Field C is the destination field and contains eleven post-decimal fields. After executing this example code, field C should contain -1.47600000000 but contains

This article discusses a case that causes a checker error and provides a method to work around the error. Problem: Compiling a program with the CREATEXFD compiler directive that uses the VAR_LENGTH XFD directive on a file description results in the following compiler error (once for each time the source includes the VAR_LENGTH directive): COBCH1508E Unknown XFD directive: '' : "test.cbl"(63,21) However, the section Using XFD Directives in your Database Connectors User's Guide states that VAR_LENGTH is a supported directive. Resolution: Install Net Express or Server Express 5.1 WS2 HF 4 or later when it becomes available. Until this version is available, however, you can use the following work-around: Change the underscore character in VAR_LENGTH to a hyphen: VAR-LENGTH. Incident Number: 2266175 Old KB# 14265

*1041-E Data item has zero size, or group not yet completed. Value of 0 assumed. * 27-S Number too large Rebuild complete with errors 01 ADDR. 05 ADDRS. 10 ADDR-LINE PIC X(20) OCCURS 6 INDEXED ADDR-IDX. 78 FIND-STRT VALUE LENGTH OF ADDR-LINE. 78 FIND-LEN VALUE LENGTH OF ADDRS - LENGTH OF ADDR-LINE. 01 ANOTHER-RECORD. Problem: *1041-E Data item has zero size, or group not yet completed. Value of 0 assumed. *  27-S Number too large Rebuild complete with errors 01 ADDR.    05  ADDRS.      10  ADDR-LINE       PIC X(20) OCCURS 6 INDEXED ADDR-IDX. 78 FIND-STRT         VALUE LENGTH OF ADDR-LINE. 78 FIND-LEN  VALUE LENGTH OF ADDRS - LENGTH OF ADDR-LINE. 01 ANOTHER-RECORD. Resolution: The 1041-E Data item has zero size, or group item is not yet completed, value of 0 assumed, and the 27-S Number too large error a

Vutil –unload encounters a 2 gigabyte limit Problem: The Vutil –unload process on a Vision index file executes until the output file hits approximately 2 GB. As an example, here is the output when this occurs: # ls -l CFSDATA* -rwxrwxrwx 1 root system 1436363264 Dec 23 10:16 CFSDATA -rwxrwxrwx 1 root system 826272256 Dec 23 10:16 CFSDATA.vix # # vutil -unload CFSDATA unCFS Record size = 2000 Unloading CFSDATA... unCFS: write error # # ls -l unCFS -rw-r--r-- 1 root system 2147483647 Dec 23 10:15 unCFS # Resolution: On operating systems that support line sequential files greater than 2 GB, setting “USE_LARGE_FILE_API” to “1” resolves the issue. Note that this must be set in the environment, since Vutil does not use a configuration file. Incident Number: 2262370 Old KB# 14205

This article explains why a table that should be loaded into syscat might not be available using AcuXDBC on a Linux operating system. Problem: A table is not available using AcuXDBC in a Linux environment. The .xfd file has been loaded, the FILE_PREFIX environment variable is correctly set, the Vision file is available, but AcuXDBC cannot see the table. Using asql.sh, this is the error: SQL (/? for help) ==> select count(*) from abcd;select count(*) from art***** ERROR: VISION: File 'abcd' does not exist Resolution: It's possible that the Vision file has an upper-case name, while the rest of the information (such as the table name inside the .xfd file) is lower-case. Check the name of the Vision file, or set this variable in the acuxdbc.cfg file: FILE_CASE UPPER # Values for File_Case are Default (case ignored), # Lower (filename converted to lower case), and# Upper (filename converted to upper case). Old KB# 14213

This article explains how to bind a compiled application with a DB2 collection ID of greater than eight characters without truncating the collection ID. Problem: Bind fails for a compiled application, and the collection ID truncates at eight characters. For example: SQL0061W The binder is in progress. SQL30104N Error in BIND or PRECOMPILE option "LANGLEVEL " with value "". SQLSTATE=56095 SQL0082C An error has occurred which has terminated processing.SQL0092N No package was created because of previous errors. SQL0091N Binding was ended with "3" errors and "0" warnings. Resolution: Compile and bind your application as follows: At a UNIX command prompt, enter the following to compile the application using the DB2(UDB-VERSION) compiler directive: DB2(UDB-VERSION=version) Where version is one of the following: Version Description V8 DB2 version 8.x V9 DB2 version 9.1 V95 DB2 version 9.5

This article describes how to configure the PowerTerm terminal emulator. Problem: Ericom's PowerTerm is a terminal emulator that is shipped with the Micro Focus Net Express product. PowerTerm is used for UNIX and LINUX environments. When executing sde or the Micro Focus license administrator, the windows are not formatted correctly or the line drawing characters are not correctly displayed. As with most emulators you have to configure it so that the line drawing, colors, function keys and overall window display are acceptable. Resolution: To configure PowerTerm: Click Terminal > Setup. On the Emulation tab, click SCO ANSI. On the Display tab, select 80 Columns and 25 Lines Per Screen. On the UNIX/LINUX side: export TERM=ansi If the environment variables COLUMNS and/or LINES are set, be sure they are set to: export COLUMNS=80export LINES=25 Old KB# 14263

Fileshare server cannot be started. The error message is “A CCINAMPU listen call failed, please re-try with fewer active connections.” Problem: The Fileshare Server cannot be started. The following error message appears: FS148-I CCI return-code (function 07) = 00000009CCINAMPU-0030E A CCINAMPU listen call failed, please re-try with fewer active connections. This error occurs only when using the “named pipes” communication protocol, that is, ccinampu. The error does not occur when using the ccitcp communication protocol. Resolution: For named pipes protocol (-cm ccinampu), Fileshare will create a FIFO in /usr/spool/uucppublic/.CCI named .servername (where .servername is the name specified when starting Fileshare). It is necessary for the directory /usr/spool/uucppublic to exist. If this directory does not exist, the error will occur. Some UNIX or Linux machines do not include a directory named /usr/spool/uucppublic. If this directory does n

When compiling the following program within the IDE, a 163 error differing from that generated from the command line results: $SET CHECKNUM IDENTIFICATION DIVISION. DATA DIVISION. WORKING-STORAGE SECTION. 01 N PIC 9(8) VALUE 0. 01 M PIC X(8) VALUE " ". PROCEDURE DIVISION. MOVE SPACE TO M. MOVE M to N. DISPLAY M. DISPLAY N. GOBACK. its error message is displayed as Object Code error : file 'a' error code: 163, pc=0, call=1, seg=0 163 Illegal character in numeric field (in D:\\anken2\\7800\\7848\\demo\\a.cbl, line 9) When the same program is built from the command line and run the message is: Object Code error : file 'a' error code: 163, pc=87, call=1, seg=0 163 Illegal character in numeric field Why is there a difference? Problem: When compiling the following program within the IDE, a 163 error differing from that generated from the command line results:       $SET CHECKNUM        IDE

When using a WSDL to generate a Web Client, an error is shown "Error: Cannot find type information" Problem: When accessing Services > Generate Client > using WSDL , the following error is received: Error: Cannot find type information for:'optInStatus' of'CPNI' of'accountSummary' of'GetAccountSummaryReturn 'The header of the WSDL being used:<?xml version="1.0" encoding="UTF-8"?><wsdl:definitions targetNamespace="http://simulation.testspace.com/omnia" xmlns:apachesoap="http://xml.apache.org/xml-soap" xmlns:impl="http://simulation.testspace.com/omnia" xmlns:intf="http://simulation.testspace.com/omnia" xmlns:soapenc="http://schemas.xmlsoap.org/soap/encoding/" xmlns:wsdl="http://schemas.xmlsoap.org/wsdl/" xmlns:wsdlsoap="http://schemas.xmlsoap.org/wsdl/soap/"xmlns:xsd="http://www.w3.org/2001/XMLSchema"><!--WS

EXAMINE WSVAR TALLYING UNTIL FIRST "Y". produces compiler error Problem: EXAMINE WSVAR TALLYING UNTIL FIRST "Y". produces compiler error Resolution: Use compiler directive dialect(OSVS) or consider using the following INSPECT statement         INSPECT WSVAR TALLYING count1 For Characters Before "Y" where count1 is declared as pic s9(9). Old KB# 14248

This article describes how to set up RUMBA to enable users to enter non-numeric characters into Mainframe numeric fields. Problem: Does RUMBA support the Numeric Override feature? Resolution: The standard 3270 host applications allow the definition of data input fields as either numeric only or alphanumeric. If a user tries to enter a non-numeric character into a numeric field, 3270 applications do not accept the input. RUMBA Mainframe display supports a Numeric Field Override feature to allow for non-numeric characters to be entered into numeric fields. By default, this option is disabled in RUMBA. Perform the following steps to enable this option: Launch a RUMBA Mainframe Display session. Click Options > Edit > General. Check Numeric Field Override; then click OK. This enables the Numeric Field Override option, and enables users to enter non-numeric characters into numeric fields. Old KB# 14235#Rumba

This article describes what C compiler to use for relinking the runtime with C libraries on Unix/Linux. Problem: You have created your own C libraries or you want to use 3rd party C libraries that will be called from ACUCOBOL-GT programs. Resolution: You need to relink the runtime by using An ANSI 89-compliant C compiler supplied by the vendor of your UNIX system, either included as part of your UNIX system or as an add-on option. Check with your operating system vendor to see if the C compiler is capable of building shared libraries. If it is not, you will need a full version of the compiler to relink the ACUCOBOL library. Please remember than Bundles will not be enough, and a full version of the compiler will be needed for most systems. Detailed information on interoperating with C and C is provided in the manual A Guide to Interoperating with ACUCOBOL-GT, Chapter 6. Incident Number: 2181911 Old KB# 14268

This article describes how to create an ACUCOBOL-GT program that attaches a Word document to an Outlook email. Problem: You want to learn how to send Outlook messages with ACUCOBOL-GT and/or you want to learn how to attach Word messages to an email. Resolution: To work with Outlook and Word, we are going to use our ActiveX interface. To do this we use the utility “axdfgen” to generate copy files describing the interface we can use to communicate with the ActiveX. We add them as copies in the Special-names section of the program. For example: SPECIAL-NAMES. COPY "word.def". COPY "outlook.def". If you need more information about creating these copy files, refer to ACUCOBOL-GT User Interface Programming Chapter 5, section 5.3 ActiveX. To use Word and Outlook you need to declare certain variables: 01 wordApp handle of Application of Word. 01 wordDoc handle of

This article describes the read part record error: "End of File before End of Record", or "file open in wrong mode". Problem: This error occurs when trying to read a control card file within a COBOL program. An example is a one-record parameter file with 9 bytes of data and filler to byte 80. Resolution: This error can occur when the file organization specified in the program does not match the actual file type. If no organization is specified in the program, the default file type for sequential files is record sequential. A file created with a non-Micro Focus file editor will default to an organization of Line Sequential. The resolution is to change the SELECT statement to add the phrase: Organization is Line Sequential Old KB# 14229

This article explains how to change the port that the Windows Data Server is using. Problem: How do I change the port that the Windows Data Server is using? Resolution: On your Data Server system, stop the Data Server. Do this from the Windows Control Panel's Services utility, available under Administrative Tools. At a command prompt, enter regedt32 to start the registry editor. Navigate to HKLM\\Software\\ODBC\\ODBC.INI\\SimbaServer\\SvrPort. Change the default port number,1583, to either 1584 or some other port number that you know is not in use. Restart the Data Server. Do one of the following: If your system has the Database Administrator version of the Data Client installed: Start the Data Client. Click the entry for the Data Server system; then click Modify. Change the port number to the same number you specified in the registry in step 4 above; then click OK. If your system does not have the Database Administrator

This article explains what you can do when you get this error. Problem: When compiling to ".exe" and attempting to link the new Database Connector EXTFH module, ACUFH.dll, this error is seen: unresolved external symbol _ACUFH. Resolution: The initial release of Database Connectors did not come with a way to link to the ACUFH.dll. Future versions will include ACUFH.lib, the import library for ACUFH.dll, allowing COBOL programs to be linked. In the meantime, use "int" or "gnt" level code for execution. It is also possible to create your own ACUFH.lib by creating a dummy ACUFH.dll with Visual Studio and having a single export (ACUFH). Incident Number: 2279117 Old KB# 14286

How can AcuXDBC Server be set up as a Windows Service? Problem: When the server is rebooted, the AcuXDBC Server is killed and must be restarted manually. How can it be configured to run as a Windows Service? Resolution: AcuXDBC Server does not currently have built-in capability to run as a Windows Service. That functionality may be added in a future release. In the meantime, here is a Microsoft KB article detailing how to set up any .exe file as a Windows Service: http://support.microsoft.com/kb/137890 Such user-defined services cannot run a batch file — it must be an .exe file. Normally, AcuXDBC Server is started from a batch file, acuxdbcs.bat, by issuing the command "acuxdbcs -start". Examining the acuxdbcs.bat file reveals that the "-start" option results in execution of the xdbcsrvr command. The details for this command are found in the section 5.7, "Starting AcuXDBC Server," of the AcuXDBC User's Gui

After my iSeries has been restarted, how can I auto start my OnWeb server? Problem: Older versions of OnWeb Server for iSeries do not have the capability to auto start. Users have to manually start OnWeb Server every time the iSeries restarts. Resolution: With OnWeb Server 7.2.5 and above, customers can edit their QSTRUP program on the iSeries to automatically start the OnWeb Server. Example code for a default installation that used the naming convention ONWEB: /* Start ONWEB Server */ ADDLIBLE LIB(ONWEB) MONMSG MSGID(CPF0000) STRSBS ONWEB MONMSG MSGID(CPF0000) OnWeb Server and the Apache Web Server will be started automatically via the addition of STRSBS ONWEB. Incident Number: 2262363 Old KB# 14216#OnWeb#Rumba

This article describes what to do when Vision files are created without permissions on Linux. Problem: Vision files are created with no permissions on a Linux server in a thin client environment. Files look like this: ---------- 1 user group 1024 Feb 10 16:49 filmografia.vix---------- 1 user group 2048 Feb 10 16:49 filmografia Resolution: Vision files created via thin client assume permissions depending on the UMASK value set inside the AcuAccess file. The default UMASK value is 002, which means that Vision files will be created in this manner: -rw-rw---- 1 user group 1024 Feb 10 16:49 filmografia.vix -rw-rw---- 1 user group 2048 Feb 10 16:49 filmografia If UMASK has been modified, persmissions will be set accordingly with the value provided. Setting UMASK to 777 causes files to be created with no permissions. Incident Number: 2266524 Old KB# 14285