What: Using djengine to run a process continuously (from the command line or as a service) eventually exhausts the system memory. The process will either terminate or cause a General Protection Fault (GPF).

Why: Modified djengine so it does not reload the process every time it is executed.

When: 2004-07-14

What: Running a transformation with -sm option set in engine does not behave correctly. When set to sm 0, a mismatch error should be displayed. When set to sm 3, a Match-By-Name behavior should be displayed instead of Match-By-Position behavior.

Why: Map SSN column to target position 4 (matching by name).

When: 2004-10-08

What: When executing a COM object that accesses transformation sample information via the ISampleInfo class, a null pointer error occurs.

Why: A message box now displays the sample type of the transformation (expected behavior).

When: 2004-12-30

What: When attempting to execute a multi-threaded process flow on AIX with max concurrent threads set to something other than 1, causes a segmentation fault (core dump).

Why: Two issues are involved here. On AIX the default stack size for executing threads is 96k, which is low when compared to other platforms. A user can set the environment variable AIXTHREAD_STK to specify a larger value from the shell that they can execute the engine from. For example, export AIXTHREAD_STK=786432 will increase the stack size to 768k or eight times larger than the default. The other problem is a conflict when accessing an object used in the processing of RIFL code when using multiple threads. This can potentially affect all platforms.

When: 2004-08-03

What: Error loading IProject interface.

Why: Need to add Xalan library to link line of Cosmos common libraries on AIX to avoid crash when loading an ip.xml project file from Java. The JVM fails to load some necessary libraries at runtime.

When: 2004-07-23

What: When loading a map or process with the SDK, if the log file contains a macro reference, this macro is interpreted and the original macro reference is lost. If the map is saved, the loss is saved to the file.

Why: The reported log file name still contains the macro reference and only interprets it when necessary (e.g. for actual file I/O).

When: 2004-07-23

What: Map Designer displays an error when resaving a map that was previously saved by the SDK (when the map is opened via Repository Explorer).

Why: Fixed. The files can be resaved and the error message no longer occurs.

When: 2004-08-02

What: When running map from Engine, it looks for map.xml file stored in tf.xml file instead of the current directory where the tf.xml file is located.

Why: The transformation uses the map.xml file stored in the tf.xml file in C:\temp\testpath.map.xml. Engine should look in the directory where tf.xml file is located. If file is not found, then look for map.xml stored in the tf.xml file.

When: 2004-11-19

What: If the install path is invalid, Map Designer displays an automation error when trying to build the connection list. No error messages are produced in DJSpoke::BuildConnList when the directory from which you are trying to build the list does not exist. Error messages are required when the Connlist cannot be built.

Why: Changed BuildConnLists so that if the installdir/connections directory does not exist, and if no factory directory is specified in the ini file or it does not exist then the current directory will be used. In this case, a warning message will be logged giving the directory (or directories) that cannot be opened and the directory being used.This will most likely result in empty Connlists, but no automation error occurs. The GUI will still produce an error message when trying to set the spoke type, but this is expected.

When: 2004-10-22

What: Some paths are being stored with unnecessary relative path symbols.

Why: Made changes so that a relative path is not constructed when two paths match only by the URL scheme and drive letter.

When: 2004-07-01

What: CreateMacro method only has one parameter for the name. The COM SDKs have three parameters (name, value, and description). The parameters are hard coded as NULL and the method only accepts one parameter.

Why: Added a three-parameter version of createMacro to IMacros and Macros. Also added setValue and setDescription methods to IMacro and Macro.

When: 2004-10-15

What: The Project.Load() method has to be synchronized when running multiple threads and connecting to a JMS queue. If it is not, then the connections for the sessions fail on all of the threads except the last one that is loaded.

Why: Added an internally synchronized load procedure and changed Load method to call it.

When: 2004-11-19

What: FileWrite() does not display error in Java SDKs.

Why: Fixed a problem with the error events not returning the error code.

When: 2004-07-14

What: When creating a package from the Repository Manager, the JAR file cannot be extracted from the CLI. However, if the package was created from the CLI there is no problem. Djar seems to have a problem creating a directory.

Why: Fixed a problem where the DJXString operator+= could reference a null DJXStringRep object.

When: 2004-09-07

What: On OS/390, reading a license file fails with a signature mismatch error.

Why: Assigning a wchar_t * to DJXString causes it to lose the encoding.This is because there was no operator=(const wchar_t *) method defined, causing a default DJXString to be constructed (with OEM encoding) and assigned.

When: 2004-07-14

What: Need Extended Unix Code (EUC) encoding for Engine. The EUC encoding systems are a group of encodings for CJK character sets. They were defined in ISO-2022 for use in 8-bit systems, The four flavors of EUC encoding in use are EUC-JP, EUC-CN, EUC-KR, and EUC-TW.

Why: Added EUC-JP, EUC-CN, EUC-KR, and EUC-TW to connector encoding property. Also added algorithm to handle 3-byte and 4-byte encoding for EUC-JP and EUC-TW in txt2cpt.cpp and fileproxy.cpp.

When: 2004-08-10

What: The SOAP API used internally by the Engine uses containers to keep track of child elements and the namespace stack. In some cases these containers are being passed by value instead of by reference causing some redundant object creation and ultimately impacting performance.

Why: Fixed issues with the handling of the containers. (Another related fix in bug 18088)

When: 2004-07-27

What: The SOAP API used internally by the Engine uses containers to keep track of the namespace stack. In some cases these containers are being passed by value instead of by reference causing some redundant object creation and ultimately impacting performance. (related to same problem in bug 18086)

Why: Rewrite NSStack class to have a separate DJObjectBase-based implementation. Pass around NSStack objects instead of actually copying strings/vectors.

When: 2004-08-27

What: A conversion is aborted on the first truncation error, even though the error is handled in the IConversionListener.OnConversionError method. For comparison, the same methods in the COM interface have a return type of HRESULT that indicates whether the map should abort or continue. The methods in the Java interface have a VOID return type. The calling code returns the error instead of an OK code, after calling the event handlers.

Why: The abortion after error occurs is controlled by native function FireConversionError. In current implementation, whenever error occurs, it fires an error and aborts the conversion. So users do not have any control over errors. DJEngine has an option called MaxErrorCount to allow users to control errors. This also seems an appropriate way for Java API to control errors.

When: 2004-09-07

What: If you have a DJAR with directory structure in the DJAR, it cannot find the map.xml file. The problem is that Repository Manager adds the directory structure. If you create the DJAR from the command line, there is no directory structure and it works correctly.

Why: The problem was occurring when parsing paths that contained both slash and backslash as separators. In this case, (on Windows), the first backslash would be found, and no search for slash would be performed.This messes up the search algorithm in the djar file, because the Java program ends up passing a complete path to the djar file using backslashes, while the path within the djar file is specified with slashes.

When: 2004-09-17

What: When running a process, the -l logfile parameter does not generate a logfile after the process is run.

Why: Changed the default parameter for the wide character version of DJGlob::StrToBool to be 0, and put the default string in the cpp file so wcfix converts it. Also switch parsing to use a wide character delimiter so wcfix converts it.

When: 2004-08-19

What: Attempting to run a djar file without specifying that it is a process causes a General Protection Fault (GPF). Running with the -pe option djengine -pe djarfile works.

Why: The problem was occurring because the Engine code treated the djar file as if it were a transformation. The code to load the transformation would then crash. Changed the Engine code so that a djar/zip file will be loaded as a process.

When: 2004-10-19

What: The dj800ec.jar file shipped with UNIX builds (AIX 8.4, Solaris 8.2 and 8.4) is missing these files:IMacro.class, IMacros.class, Macro.class, Macros.class, NMacro.class, NMacros.class, QMSG_ACTION_TYPE.class

Why: Added missing classes to both dj800ec.jar and libdj800ecjni.so. This issue applies to all UNIX platforms and OS390.

When: 2004-08-23

What: When trying to run a djs file on OS390, the load fails with the simple message "killed".

Why: Problems were being caused by char/wchar_t comparisons, either because of character constants or places that wcfix could not modify. Fixed these occurrences.

When: 2004-08-20

What: Using the Java SDK, we have a class that runs a simple process. Within this process, we are trying to write to a file that does not exist. The Java application returns an Error 28029 "Unspecified Error" when it should return the message "Expression step error encountered."

Why: Added code to set the LastError message to "Expression step error encountered" if the expression execution fails.

When: 2004-08-30

What: The Bin Function is not available in the OS/390 Engine.

Why: Changed code to do a linear search through the keyword list on OS390, since the EBCDIC ordering is different than ASCII and the keyword list therefore is not sorted on OS/390.

When: 2004-08-12

What: The InitCaps Function does not work.

Why: The function that performs this was using isalpha on Unicode strings. Changed to use CCISWALPHA.

When: 2004-08-19

What: When running the -PF switch on the CLI for the OS/390 engine, the resulting output file is unreadable.

Why: Ensure that all strings get written to the Profiler output file encoded as UTF-8.

When: 2004-08-20

What: Events do not fire in the correct order when running a test map with the OS/390 Engine. There is a schema mismatch warning in the log, but there is no mismatch. The Engine may not be running the map correctly as a result.

Why: The Null Connector allows setting field types, and this is done based on char values as compared to ASCII values. Changed to store/set the field types based on wchar_t instead.

When: 2004-08-18

What: When executing tf file, that connects to SQL Server 2000 Multimode and uses a DJImport object that connects to SQL Server 2000 from Java, the Java application returns an EXCEPTION_ACCESS_Violation.

Why: The COM was getting uninitialized before it should. In the GUIs and djengine, an extra call is made to CoInitializeEx that keeps COM active. This extra call is not present when accessing the Engine through Java.Changed the ComManager class to keep track of the number of times Initialize and UnInitialize get called. It will call CoInitializeEx function when the count goes from 0 to 1, and call CoUninitialize function when the count goes from 1 to 0.

When: 2004-09-20

What: Cannot run djar files on Solaris or Windows.

Why: To run a djar file, you need to specify the -pe flag that works on both Windows and Solaris.

When: 2005-02-09

What: Running two conversions parallel on multiple threads does not show performance gain on multiple CPU machines running AIX.

Why: The AIX ACE_Manual_Event seems to be in a signaled state. Insert a short 1ms sleep before waiting on the ACE_Manual_Event object on AIX.

When: 2005-01-14

What: Add "Pervasive PostgreSQL" sc.xml/tc.xml files to reference the PostgreSQL Connector.

Why: Added tc.xml and sc.xml files.

When: 2005-01-13

What: Remove minor version number check from the licensing code.

Why: Commented out minor version number in the licensing code.

When: 2005-01-12

What: Memory issues (both leaks and initial usage) could be related to access violation.

Why: Fixed the access violation problem.

When: 2005-01-14

What: Total runtime for the process reports from multiple CPUs and not the total time to run.

Why: Modified the method that keeps track of elapsed time for a process.

When: 2005-01-07

What: Memory leak due to unnamed compiler objects.

Why: Fixed the problem where DJBasicCompiler objects were created without names, thus causing generation of unique names for each instance.

When: 2004-12-08

What: Performance problems found on OS/390 engine.

Why: Changed the code so that seek is within the buffer (and is not relative to the end), and the buffer is not flushed, then we can reset the pointers and return it without refilling the buffer.

When: 2004-12-07

What: No error message is displayed when loading a map into a Project object.

Why: Throws an exception when loading an "old" project document. This causes the Java process load routine to throw an exception (though the exception message states "Unspecified Error").

When: 2004-12-06

What: QSessions.GetItem(x) causes application error.

Why: Fixed the problem by retrieving a queue session from the Queue Sessions interface.

When: 2004-12-02

What: Loading a bad djar causes a crash/core dump.

Why: If a top-level element of process XML is loaded, but not processed it displays a DJXMLException rather than calling XMLStore::startElement.

When: 2004-11-30

What: When loading a .djs file using the 8.x CAPI, you get an unlicensed Connector error if you run it.

Why: In DJSetInitializationFile, read the license file (and check licensing).

When: 2004-11-22

What: When you have the name of an Item and try to assign the results to the variable GetItemByName() function, it doesn’t work.

Why: Fixed an access violation caused by an uninitialized variable. GetItemByName() now returns an error.

When: 2004-11-18

What: Execution time is not computed correctly when executing from Map Designer.

Why: Stop the elapsed time counter before resetting the connect string and reconnect it to the source.

When: 2004-11-10

What: Problem with Integration Server Engine Log tf.xml execution.

Why: Modified the COM and Java APIs to make sure the log file is open on the same thread when the conversion is executed.

When: 2004-12-04

What: When trying to run djengine by passing the name of a .map.xml file instead of tf.xml file, it causes an access violation.

Why: In sepfile, make sure that the filename and extensions are initialized to an empty string.

When: 2004-11-23

What: When running an old djs file from the command line and supplying a new log file location/name, the log file is not correctly overridden. The Engine writes the log information to the log file specified in the djs file.

Why: Changed djs loading code to OverriddenLogfile flag.

When: 2004-11-04

What: The JDK at times installs/sets registry keys under HKLM/Software/JavaSoft/Java development kit, but not under HKLM/Software/JavaSoft/Java Runtime Environment.

Why: When finding the JRE through JDK home registry key, be sure to append "\jre" to the Java home. Provide error messages in the JVM Manager class when you have logging problems in RIFL Java object access.

When: 2004-10-29

What: The Java application loads a process and then executes the loaded process multiple times. Prior to calling Project.Run (Boolean) the SharedExpressionContext interface sets variables that are used within the process for the Source and Target file locations. The application successfully completes several hundred iterations prior to an exception due to the corrupted variables.

Why: This process ends up generating more than 65535 compiled strings. Changed the compiler and executor to reference string indexes as "longs" rather than "shorts".

When: 2004-10-27

What: The COM GetItem() call displays an unidentified error.

Why: Fixed an access violation caused by an uninitialized variable.

When: 2004-11-11

What: DJEngine service crashes with an application error when adding a file to a file folder queue within a process.

Why: Changed SetGlobalVariables so that it does not reset the program strings. Changed compiler objects to take on the name of the step to avoid adding to the program strings with automatically generated compiler names.

When: 2004-11-08

What: When creating a conversion object, call Load to load a tf.xml, and run it. Load() reports no errors even if the file is not existing. Run() reports "Connector type invalid or unspecified".

Why: Changed the code. In case of an error, the fireConversionLoaded initiates before ECReportError and shows an exception.

When: 2004-11-16

What: Segmentation fault with AIX Engine.

Why: The process runs out of memory. The linker flag -bmaxdata:0x40000000 to set the soft ulimit to 1GB is not applied to the executables, but only to the shared libraries. Made changes to apply it to djengine, and also changed from 1GB to 2GB to avoid the "out of memory" condition.

When: 2004-10-22

What: When running the process on Solaris, the Engine displays an error message," Bus Error (coredump)".

Why: Changed SZ_FNAME to 512 whenever it is not defined by the platform. Also added a couple of checks that results in an error rather than a coredump/General Protection Fault (GPF). The immediate solution is to edit the ip.xml file and fix the path strings.

When: 2004-10-27

What: The Visual Basic application crashes with an error message "Memory Can't be Read".

Why: The get_Item method for ConversionOptions and InterfaceOptions check the index value for validity, but they do not create the corresponding option object. Changed code to create the corresponding option object.

When: 2004-11-19

What: When running multiple threads, the Process occasionally displays an error message," 09/29/2003 10:28:59 4 0 O Global At line 5 of module28: Type Mismatch (Unknown queue type)".

Why: The flag indicates that the queue connection list is built. It is set before the list is actually built. This allows another thread to look up a type in the list before the list has any entries. Changed the code so that the flag is not set until the list is built.

When: 2004-10-26

What: ISharedExpressionContext interface contains methods for creating and getting messages, but no methods for removing messages.

Why: Added removeMessage methods to SharedExpressionContext.

When: 2004-10-11

What: In the latest 8.2.12 Solaris build, the garbage characters are written to the queue in the message body after the contents of the XML structure.

Why: The contents of the queue message body are passed directly to NewStringUTF. However, NewStringUTF expects a null-terminated string, and the content of the queuemessage body is not necessarily terminated by null. Copy the contents to the DJString, and pass it to NewStringUTF.

When: 2004-09-23

What: When the map encounters a negative number in the PSTNG_AM column, the map crashes with an "IOT/Abort trap (coredump)" error.

Why: On AIX (OS/390 engine) characters by default are unsigned, but the baseone code expects them to be signed. For AIX, use the B1INTCAST as modified for OS/390.

When: 2004-09-29

What: Visual Basic crashes when running v7.55 DJS file in C-API with v8.0 engine.

Why: When checking for an overridden log file in the convert control verify that the convert control is set.

When: 2004-11-22

What: Java application hangs after the Engine is initiated.

Why: Call Disconnect() in destructor to make sure all threads started by the JMS provider are stopped.

When: 2004-09-15

What: When running a djar in the Java SDKs on Solaris, it crashes.

Why: Changed DJZipFileImpl management from a single linked list that is protected by a "mutex" to a linked list per thread. This ensures that a ZipFileImpl object is not shared between the threads.

When: 2004-09-22

What: We need to separate the three main designers into separate components: Map Designer, Source View Designer, and Structured Schema Designer, so that they are separately installed via setup.ini. For each designer, if it is set to false, do not install the EXE file and exe and do not create a shortcut on the Start menu.

Why: Examined the settings in the ini file where all the separate products are set to true or false.

When: 2004-11-09.

What: A Message object can have a set of typed properties. The current Message interface has no method to let a client get the types of the properties. As a result, if you try to call the method getIntProperty(name) on a double property, incorrect values are returned to the client. You need to add method getPropertyType(name) to the Message interface.

Why: Added method getPropertyType(name) to the Message interface.

When: 2004-11-29

What: Receive an error message when running a process validating 270 HIPAA file.

Why: The validation problem is fixed for this process.

When: 2004-08-25

What: When running the EDIFACT test map on the OS/390 Engine it displays an error message "Map is unable to connect to the source."

Why: Added new Shared libraries so that the EDIFACT map runs fine.

When: 2004-08-20

What: When you run the EDI testmap on OS/390 Integration Engine. The following error message is displayed:Parser cannot fully parse the edi file src/dummy_824_3040_src.edi due to exceeding the max count of segment in schema.

Why: Make changes to get the EDI connector working on an EBCDIC system. Note that the source file must be in EBCDIC, to match the target side, which always writes an EBCDIC file.

When: 2004-08-20

What: PostgreSQL installs the new version of openssl in the system32 directory. The new version has deprecated one of the entry points used by the Engine.

Why: Replaced call to OpenSSL_add_all_algorithms() with OpenSSL_add_all_ciphers() and OpenSSL_add_all_digests() in the DJZipFile.

When: 2005-02-04

What: If you install Cosmos 8.2.0 and open the sample process, FileList.ip.xml and run it, Process Designer closes itself.

Why: Narrowed it to the FileList() function that is causing the sample to crash. If it is commented out, it runs without incident.

When: 2004-09-30

What: Depending on which databases you write to different errors are displayed and DJExport does not work.

Why: The problem occurs because the datatypes do not get remapped into something that the spoke understands. Changed the code in the generic SQL export _OpenOutput method to check for a connection to the database.

When: 2004-08-06

What: Incore lookup functions called with user defined functions as parameters, return zero instead of a default value when the lookup does not find the passed in value.

Why: This problem occurs in cases where the function being called as a parameter is returning the last of its own parameters, and the first parameter to this function is being passed by reference. In this case, the reference may be to another operand on the local stack, which at the end of the function call sequence has all values set to 0 (to ensure objects are released).The solution is to check the return value for an opref, and if found, resolve the reference to push the actual value.

When: 2004-07-29

What: Calling a User Defined Function within another User Defined Function causes the second UDF to be called recursively until a Runtime Error occurs. This occurs when calling the UDF from outside the scope of the Function definition.

Why: The problem is being caused by the way a user-defined function is referenced in the generated byte code. This was being done by putting in a small ordinal for the function, used to look up its address. Changed the code so that code refers to the function's definition object and uses it to look up its address within the current module.Also changed the definition object itself to save references to functions called from within it so that the compiler can reference the code of these functions.

When: 2004-11-16

What: Duplicate function definition error in code module. This happens when two functions are trying to call a function. And both functions calling the function occur before the called function in the code module.

Why: The problem was that the lexer was not checking the list of forward declarations. Because of this, the identifier was not recognized as a user-defined function the second time around and an attempt made to generate another forward declaration, which caused the error.Changed code so that when checking if an identifier is a user-defined function name, it checks the list of forward declarations as well as every other place.

When: 2004-11-16

What: Trying to compile a RIFL script with a forward reference to a user-defined function fails with an undefined function error.

Why: In the case of a forward function, a new function object is created at definition time, leaving the function object created for the forward reference undefined, resulting in the error.Changed function generation code to not create a new function object if one is found.

When: 2004-08-09

What: Incore lookup appears to be returning values for lookup values or key values that are not in the lookup file. It looks like duplicates from earlier records.

Why: Data type conversion values need to be compared. The keys are currently stored with the data type of the first key added, and values being looked up are converted to the appropriate data type. There may still be cases where this will fail (i.e. the keys are strings, such as "0050", and a lookup value of type LONG with a value of 50 will still fail).

When: 2004-10-19

What: In the RIFL script editor, the last Japanese character is truncated.

Why: The broken Japanese character display in the RIFL script editor is fixed.

When: 2005-01-21

What: RIFL is not working internally in Unicode as expected.

Why: Changed the code so that the length of the Unicode contents is retrieved, thus giving a length of 5, which is the correct value in this case.

When: 2005-01-10

What: The RIFL script that has an error in syntax displays an error dialog. It is expressed as error check at the time of execution and displays a part of Japanese character sequence in a dialog that is either missing/truncated.

Why: Fixed the truncated Kanji character.

When: 2004-12-22

What: Opcodes are truncated for a couple of functions when expressions are optimized in method DJBasicCompiler::RegenerateProgram.

Why: Fixed sizeof(short) to sizeof(long).

When: 2004-12-09

What: Automation error is displayed when the Flow Control function is invoked in RIFL Editor.

Why: Add a "before" call to CodeMax1.SetCaretPos in frmExprBld.InsertToken.

When: 2004-12-08

What: The RIFL date conversion by the Format function with Japanese extension causes errors and the results are not displayed as expected.

Why: The errors are caused by the new implementation of datetime for 8.0 while Cosmos J still uses 7.55 code.

When: 2004-12-08

What: An error occurs when using a DJX statement in a SQL Step in Process Designer. There is a space between DJX and SQL statement in the step.

Why: Allow spaces between DJX and the following SQL statement.

When: 2004-11-17

What: "For each" is supposed to iterate through each member of a collection and return the items one at a time. Thus in an array of numbers, it should return each number (in unknown order).

Why: The initialization for the DJForArray iterator object sets the dimension count to 3, which causes indexing of array (index validation checks) to fail. Reset it to vai.size().

When: 2004-10-22

What: Application returns a memory access violation (General Protection Fault (GPF) when loading a single dimensional array.

Why: Fixed a typo that caused a GPF violation when accessing a single dimensional array.

When: 2004-10-05

What: Resetlookup function does not work if the path to tlookup file uses macro, $(macro). Using fileappend to add a new value to the tlookup file. After the fileappend function, tried using the resetlookup function to reread the file, but does not work when using the macrovalue function.

Why: Changed lookup management code to expand macros.

When: 2004-10-21

What: A transformation that is run repeatedly from within a Java program that creates and then destroys Java objects generates a memory leak.

Why: DJ_JVMInfo objects were not being deleted when they were no longer referenced. Changed code to get the refcount, and if it is 0 (and the object is not the static version) delete it.

When: 2005-01-18

What: If a forward reference is made to a function that is defined later, it is either implicitly or explicitly defined as global (public or private), an undefined function error occurs.

Why: When looking up a function definition and before defining it, look at the modules list of functions. If found and no code is associated with it, assume that it is a forward reference and use it. This allows forward referencing a function that is later found to be global.

When: 2004-09-07

What: Need new Btrieve v8 DDF reader.

Why: The adaptor was created and added.

When: 2004-07-14

What: Validator generates a warning when a mandatory element with a numeric data type contains only a zero.

Why: Leading zero in a numeric value is now referred to the numeric value which contains zero digits before any non-zero digits. Therefore, the values containing zero digits such as 0 and 000 will be valid. Values such as 00345 and 010 are invalid.

When: 2004-07-29

What: When an element has multiple code values specified the Like operator is used and a "|" (pipe) delimited string that contains all the valid codes for the element.Validation works fine, except since the operator is Like, a value of "0" or a value of "1" is considered a valid value.

Why: Implement IsLike function to evaluate the Like operator instead of DJ Regexp. Regexp may not correctly evaluate the Like operator.

When: 2004-07-29

What: Three loops in EDI X12 856 schema have a count of zero, but the UsrRequirement is optional. The loops are MAN, SDQ, and GF. When validating the file against schema with djvalidator or Map Designer, it displays an error message "Loop maxuse exceeded".

Why: If loops have a count of zero, Document Schema Designer does not generate segments for the loop in a sample file.

When: 2004-11-30

What: In SWIFT 940 message, look at 61: lines [B.4!R.Group1]61_6 recordtypes. Using 940 schemas as an input, you cannot get to field 61.10. This is because the field 61.8 is labeled as 16x in the schema. However, the parse does not end at "// (61.9)" because it is of type x.

Why: Based on SWIFT standard, 'x'datatype can be of any character including '/', because '/' is used for the field separator in SWIFT, 'x'datatype should exclude '/' as a permitted character.

When: 2004-11-30

What: If a map that uses RIFL code modules is copied to a different machine and the map file is saved on the new machine, xmldb URLs are used to store the RIFL file locations. When the file is reloaded, the Map Designer fails to locate the RIFL files. Since RIFL files are not XML, RIFL file locations should be stored with regular Windows or UNIX paths.

Why: If the code module reference becomes an xmldb:ref URL, convert it to a file system path.

When: 2004-10-19

What: When deleting records from a MySQL table, records are actually updated.

Why: The code was deleting the record and then re-inserting it (with the changed data). Changed the code so that if it is in delete mode it does not re-insert the record.

When: 2004-09-21

What: Process Designer sessions do not retain MySQL ODBC passwords when a new map is created inside a process using the MySQL ODBC connection. The password can be entered into the session information and is saved, but it seems that the password is not being read from the ODBC administrator.

Why: The process sessions are now reading and saving password information.

When: 2005-01-21

What: Date/time fields in Excel only appear as date in Map Designer.

Why: The code checked the string for 'm/d/yy' and assumed a date before checking that the whole string was 'm/d/yy hh:mm'. The code did not check for 'm/d/yy hh:mm AM/PM'. The ordering is changed.

When: 2004-09-21

What: HL7 rules in the Standard states that if a segment is present in a message but not expected (not in the schema), it should be ignored. The message should still be able to be loaded into a map.

Why: If extra segment in data file is not expected in schema, the HL7 parser skips the segment and logs it as a warning message such as "Skip Record ID=?, Filepos=?"

When: 2004-10-11

What: The HL7 Standard states that extra components and subcomponents appearing in a message that is not expected should be ignored.In a Map with HL7 A04 as a source with extra components and subcomponents in the MSH.9 and EVN.1 fields, if you look at the Source Data Browser after connecting, you can see that the extra components and subcomponents are not ignored. The "^" is the component separator and the "&" is the subcomponent separator.

Why: If extra components and subcomponents in data file are not expected in the schema file, they will be ignored. If the connector property validation=true, it will be logged as Skip extra (sub) components in Field? in Segment?

When: 2004-10-11

What: The HL7 standard states that extra repetitions of a field that are present but were not expected should be ignored. The Map uses a source file with extra repetitions of field MSH.7. The schema states that this field cannot be repeated, yet in the contents of the source in the Map all repetitions are present. The repetition character is the tilde "~".

Why: If extra repetitions of a field were not expected in schema, they are ignored. If validation=true it will be logged as Skip extra repetitions of Field? in Segment ?

When: 2004-10-11

What: The HL7 standard states that if a segment of field is expected but not present, it should be treated as present but empty.The map uses a schema that has a required segment (EVN) and a required field in the segment (EVN.1). The test file has no EVN segment. When the Source Data Browser is opened in the map and you look at the EVN segment, a dialog box states, "There are no records of this type in your file." It seems like HL7 is expecting something like an empty dummy segment if this happens.

Why: If a segment of field is expected in schema but was not found in the data file, a dummy segment is generated with empty data. A warning message will be logged as "Add dummy record for Segment ?"

When: 2004-10-11

What: When connecting to a Paradox target, the target table name disappears after connecting to it.

Why: Set the described_table in Describe.

When: 2004-11-18

What: If a time is set as "00:00:00", it is inserted into DB2 table as a null value.

Why: The code to assign a DJTime object to a DJTimeField was checking both that the DJTime object was valid, and that it had a seconds count greater than 0. In fact, a second count of 0 is valid, and should not have been causing the field to be marked as blank.

When: 2004-11-09

What: Create a map in Map Designer that uses the Salesforce.com SOAP connector and leave it open for longer than 30 minutes after you connect to Salesforce. An error is generated saying "Insert: error soapenv:Server (INVALID_SESSION_ID: Invalid Session ID found in SessionHeader) in Form_Activate(rnstatus.frm) (25508)".

Why: Before each non-login call, check the time for which the session has been active. If it approaches a percentage of the session timeout (2 hours), relogin first to establish a new session and timeout. Queries require at least 90 percent of outstanding session time; for "queryMore" and "sending" (update/delete/insert) require 50 percent of outstanding session time; for other operations (describeGlobal/describeSObject) you require at least 10 percent of the session time remaining. Also reset the default batch size to 200.

When: 2004-11-09

What: Money data type is rounded in SQL Server 2000 and SQL Server 2000 Mass Insert. Browsers and Results column shows correct value but Query Analyzer shows the value as rounded down and the insert SQL statement from Map Designer shows a rounded down value going to the table. Multimode does correct Insert statement.

Why: When setting up a SQL Server money field, get the data from the field as a clsNumDbFld, multiply by 10000 and assign to DJQuadWord. Then get the raw data from that to bind. Also fix DJQuadWord conversion from clsNumDbFld.

When: 2004-07-02

What: If a user creates a Btrieve v8 .dat file as a target file and they fill in the password field the .dat file is created. But if the user tries to load that file as a source, it does not require the password to access the file. You may feel that it does not need the password for Read access. But if a file is created from Pervasive PSQL V8 Btrieve Maintenance Utility, or the password (OwnerName) is applied to the file from that utility, then it is required to access the data in the file.

Why: Changed create code so that after creating the file, if the user has entered an owner the file closes, reopens and then and set the owner.

When: 2004-07-02

What: Map Designer does not read all data from Sage line database. If an empty table is selected as source you will receive an automation error when connecting. Also if you select the table in the Data Browser.

Why: The problem is that the Sage driver does not return a maximum precision value for certain data types when building the data type list. Changed the code to detect the Sage driver (look for dbms name "Sterling") and if that is the driver and the max precision is 0, set it to something reasonable.

When: 2004-11-09

What: Multirecord ASCII Delimited displays 25517 error at runtime when a target file exists.

Why: The schema resolution code does not consider that the mismatch is detected; this is because the map has more record layouts than the spoke.Added code so that if this is the case and the spoke supports multiple record layouts, and the layout mismatch option is set to use map, then push the map schema on the spoke.

When: 2004-10-12

What: BIGINT data type from DB2 is converted to a "qd" on the target.

Why: Use 'lld' instead of 'qd' for sprintf format modifier on UNIX.

When: 2004-08-16

What: SQL Server 2000 big integer data type strips off the negative sign on large negative numbers. It also gives inaccurate output if the negative number exceeds 9 digits.

Why: Converting to a DJQuadWord lost the negative sign and the browser returned incorrect data. Correct data is now being written to SQL Server.

When: 2004-08-19

What: If a user runs a transformation with a Binary Line Sequential Source on Linux, a segmentation fault occurs.

Why: Changed code to use CString::begin() to manipulate the buffer, rather than c_str() on Linux. Changed compilation options to agree with the rest of the code (in particular, remove the -fomit-stack-frame option so GDB will be usable).

When: 2004-08-31

What: When target property AutoTypeTextFields is set to false, target fields with no value are written with data from the previous field.

Why: A buffer was not getting cleared before formatting cell contents. Changed code to clear this buffer before each cell.

When: 2004-08-06

What: Macro placed in SQL log property does not produce SQL log and records errors in Map Designer error log.

Why: Support the use of a macro in SQL log file specification.

When: 2004-10-19

What: A COBOL copybook containing an "OCCURS 0 TO 1 TIMES DEPENDING ON" clause will not correctly process rows where the structure occurs 0 times. Changing it to "0 to 2 TIMES" allow it to be processed as expected.

Why: Changed code that generates unique field names even when the "OCCURS maximum" is "1"so that the DependingOn attribute is processed. Generate a new/unique name only if the "OCCURS maximum" is set to more than "1".

When: 2004-09-08

What: When running a map, you receive field type mismatch messages in the log file about some source record fields.

Why: When promoting a field from text to record type, only the data type is changed. This caused a warning when mapping fields as it compared field types as well as data types.Changed the XML code to set the appropriate field type when changing data types.

When: 2004-10-19

What: Receive an automation error when a .DDF file has been applied as layout for a Btrieve source file.

Why: The problem occurs when a very large double number (>1.0e+55) is assigned to clsNumDbFld and an exception is thrown. The solution is to catch this error and reformat the number by using sprintf.

When: 2004-09-08

What: Salesforce does not update the target database.

Why: Sometimes the code changes the key values to lowercase. This causes problems when the key field is an ID field, because while generating the keys the value is lowercase, but when comparing the keys it is not. Changed code to preserve the original case of the key values.

When: 2005-01-14

What: Need to add Oracle 10g Multimode .TC.XML file that point to the Oracle 9i Multimode. The connector displays Oracle 10g Multimode on the Target tab.

Why: Changed spokenm to call the Connector "Oracle Multimode". Changed Oracle 9i Multimode.tc.xml to reference "Oracle Multimode" as both the spoketypename and basename. Added 10g Multimode with the same conditions.

When: 2005-01-12

What: When using MSSQL 2000 to SWIFT with decimals on the source and "d" data type on the target the zeros after the decimal is truncated. In 7.5.5 it displays all the zeros.

Why: SWIFT spoke handles numeric fields as text fields and values and relies on receiving the formatted number. In this case, the RIFL optimization is turned off. With RIFL optimization turned on, field to field mapping (in this case, from a decimal source field to a text target field) causes the source field to format the number to text. With RIFL optimization turned off, the actual decimal value from the source field is brought into the RIFL code, which then has to format it for the target field. This is formatted via "general" formatting, which strips the trailing zero decimal digits.

When: 2004-12-10

What: Empty elements return empty strings.

Why: Use Record BlankFields() before assigning data to fields.

When: 2004-12-06

What: Add Salesforce v5 tc.xml and sc.xml files for Salesforce.com v 5.0 Connector type.

Why: Added Salesforce sc.xml and tc.xml files.

When: 2004-11-24

What: SQL Passthrough statement with DJX does not work with Pervasive PSQL V8 connector.

Why: When resetting the cursor, ensure that the query statement is processed as SQLString.

When: 2004-11-23

What: OnConstraint event handler does not initiate for DB2 ODBC.

Why: Delegate GetNumericSQLState to connection, if there is a connection.Check for constraint errors.

When: 2004-11-08

What: Reject records are written without data when the source is XML.

Why: Set up the records in the reject proxy before pushing the records onto the real reject export.

When: 2004-11-09

What: 'Spoketype' is the INVISIBLE connector properties for EDI/HIPAA. This option is not found in TransformationOptions list in tf.xml file. This option should be saved into the tf.xml file.

Why: Modified the save code to save connector options with the OPTION_INVISIBLE flag set unless they also have the OPTION_NOSAVE flag set.

When: 2004-10-22

What: IdentityInsert property is missing in the Access Connector type.

Why: Added property to the export and general lists.

When: 2004-10-21

What: Unable to connect to the message object as a source in the map that contains XML. The internal DTD reference displays an error message in the log file.

Why: Make sure that the xmlbrowser parser is set to ignore any external DTD subset if the first try fails. Changed the actual data parser from trying to read the external DTD subset.

When: 2004-10-22

What: The default batch size for Salesforce connections is too low (10).

Why: Removed the MaxBatch connector option for the Salesforce Connectors. The value is now set to 200.

When: 2004-10-12

What: Process Designer hangs when running a map that contains Lotus Notes connection.

Why: Modified Lotus Notes connector to make sure that sessions get terminated on the same thread they are started on.

When: 2004-10-13

What: Target fields are padded with spaces when using HIPAA adapter.

Why: Now the handling of the specified padded character is correct; specify a space and you get padding, or specify none and you get none.

When: 2004-10-21.

What: Running a map with a Unicode Delimited UTF-8 target gives correct header on first run, but puts the last line of data as the header on subsequent runs if map is not shut down.

Why: The output buffer was not getting cleared in _OpenOutput. This problem was discovered and fixed in the trunk code line.

When: 2004-11-09

What: Upgraded HIPAA map gives 25515 errors.

Why: Fixed code to add EDI, EDIFACT, HIPAA, and TRADACOMS to the list of schemasInferred spokes.

When: 2004-10-06

What: When attempting to read a table set up for testing the data profiler, an error message is displayed "Numeric value out of range error" messages.

Why: When binding columns for input the p_signed attribute could be wrong. Changed code so that TinyIntParam, ShortParam, and LongParam takes issigned parameter to ensure the attribute is set correctly.

When: 2004-10-21

What: Data from Salesforce is truncated at the ampersand (&) character, if it is one. For example, in the Account table a field called Name has a value "This is a & Test". When the Map Designer reads this record, it gets "This is a ".

Why: The characters() method in StringSerializer does not append data from the subsequent calls to the method, only the first chunk of the string appears. The problem is fixed to append data from additional calls.

When: 2004-10-21

What: Specified a log message to occur when onMismatch event handler is fired when updating a table. But when no match is found in the table, no messages are written in the log.

Why: Changed code so that SQL_NO_DATA when performing a update/delete function returns B_MISMATCH.

When: 2004-10-06

What: Unable to load and upgrade .djs files that contain connectors which are no longer in the connector lists (no .SC.XML or .TC.XML files present).

Why: Changed connlist building code to check both the specified connections directory, and a deprecated subdirectory off of it. The deprecated entries will be in the connlist, but will not appear in the map designer dropdown lists because that is build simply by examining the files in the connections directory, and never recurses into the deprecated subdirectory.

When: 2004-10-21

What: In Map Designer, numbers are not shown after the eighth digit decimal. Instead, the numbers are replaced with zeroes.

Why: Changed the general decimal precision scale from 64,8 to 78,20.

When: 2004-10-05

What: Macro does not work with QueryFile.

Why: Changed control code to disconnect only if both the filename passed and the filename after macro expansion is different from the current set value.

When: 2004-09-20

What: Null error when deleting the record.

Why: Changed code to bind NULL fields to NULL values if the mode is DELETE.

When: 2004-09-08

What: If the HL7 connector uses file with multi-MSH, it fails to parse the file.

Why: If a HL7 data file contains multi-MSHs, HL7 parser does not stop when it finds another MSH. Instead, it reuses HL7 schema definition to continue parsing HL7 message until the end of file.

When: 2004-10-27

Problem: When using the "parse XML or HTML data" option from right-clicking after loading data, the following error message is displayed: "Control datlfields not found".

Solution: Changed the code to fix the problem with HTML\XML parse. The application now parses without any errors.Date

Fixed: 2005-02-10

Problem: When using the EDI parse option with Extract Schema Designer, the following error message is displayed when it comes across a line that has no data: "Invalid procedure call or argument (5)".

Solution: Changed code to get autoparse to skip over blank lines and proceed further rather than stop, because it could not add a line style.Date

Fixed: 2005-02-04

Problem: If you perform an XML/HTML Parse on an XML file, where a line has data within tags that start and end on the same line, the parse will fail with the following error: "Control 'datlfields' not found (730) ".

Solution: Changed the code to fix the problem with HTML\XML parse.Date

Fixed: 2005-02-04

Problem: The documentation states that the XML parse will create Data Fields, but the application does not.

Solution: The automated parse does not define fields unless the Begin and End tags are on the same line. Changed the code to fix the problem with HTML/XML parse.Date

Fixed: 2005-02-08

Problem: If you have a source file opened in Extract Schema Designer with multiple Accept records and try to browse the data; the Accept Record Definition dialog is displayed. If you select the "Select All" button, the "Show Fields" button text gets truncated.

Solution: Adjusted the window sizing so that it is slightly wider, thus providing enough space for the button text.Date

Fixed: 2005-02-04

Problem: The Continuation Rule in a field definition is 'Fixed number of lines'. Number of fixed lines is 3. When 'Start Rule' and 'End Rule' are floating tags, the tags for the second and third continuation lines can be seen in the data.

Solution: The application assumes you have ‘Until End Tag Found’ as your Continuation Rule, if you use a floating tag end rule. The code has been changed to address this issue.Date

Fixed: 2005-02-07

Problem: Disable the Validate button if an "open" Extract Schema does not exist. Without an open Extract Schema, the validate button is enabled and when you click on the button, it displays an error stating that the accept record does not exist. This is accurate.

Solution: A typo in the name of the Validate button caused the code that disables the button to fail. Fixed the typo and also added code to disable the Text Search and Clear Accept Records buttons, which should not be enabled when there is no script selected.Date

Fixed: 2005-02-04

Problem: The "Clear All Accept Record Fields" icon is active even if there is no open Extract Schema. Click the icon and it looks like the application is clearing the Accept Records when there is nothing to clear.

Solution: Changed code to disable the icon until there is an open schema.Date

Fixed: 2004-01-24

Problem: In Extract Schema Designer, if you set the start line in the Source Options > Display Choices to 0 (zero), and set the End Line to 50 (or any number greater than 3), and click OK. When you reopen Source Options, the last line is set to 48 (or two lower than the number specified as End Line).

Solution: Adjusted the starting line code to handle 0 (zero).Date

Fixed: 2005-02-04

Problem: If you create a Source View from Source View Designer, two files are created. One is Source Properties and the other is Source Record Layout. If you select the Source View file in Repository Explorer, right-click and select CVS > Add, only the .sv.xml file is added. The svf.xml file is ignored.

Solution: Changed code to ensure that all files associated with the Source View is added and committed.Date

Fixed: 2005-01-13

Problem: Testing a djar file on Windows works without issues, but porting the djar file to AIX causes an error.

Solution: The error message indicates that the process was run on the command line without specifying the -pe option. Changes have been made to the code so that you do not have to specify the -pe option anymore.Date

Fixed: 2005-02-15

Problem: On a Solaris platform, trying to load the Java 1.4 Runtime in Cosmos causes a segfault error to occur. The same code works fine if Java 1.3 Runtime is used.

Solution: With Java 1.3, Sun Microsystems put a symlink in the .../lib/sparc directory pointing to client/libjvm.so. With 1.4, Sun Microsystems changed this to be a copy of the file. This caused the libjava.so library to load a new instance of libjvm from the one in the sparc directory, while the application loaded the one in the sparc/client directory.Changed the dlopen call to include the RTLD_GLOBAL flag, so that only one instance of the libjvm.so gets loaded.Date

Fixed: 2005-01-27

Problem: The name v8_FullInstall.pdf appears in the "Installing Integration Engine on Windows 2000/XP" section in the Engine book.The name should be install_v8.pdf.

Solution: This issue has been fixed in the documentation.Date

Fixed: 2005-01-18

Problem: The problem resides in the fact that the "RETURN_CODE_

Solution: Fixed a problem with tracing execution paths when a step has IgnoreError set.Date

Fixed: 2005-02-07

Problem: Transformations that use a query file for the source will not run in a process.

Solution: Fixed a problem where transformations that use a source query file are prevented from executing.Date

Fixed: 2005-02-09

Problem: If the macro value used to reference a code module does not match a macro defined in the macrodef.xml, Map Designer ignores the value when the map is opened. The problem is that there is no warning.

Solution: When trying to resolve a reference from map.xml, if the file name contained macro references, one of the filenames is blank. Combined with relative path attempts, this will result in the name "./" being tested, which will exist.Changed code so that if the relative filename does not exist, it is not used in the locate attempt. This way, the file is actually not found and the original name remains in the map. During run time the missing code module will be detected.Date

Fixed: 2005-02-01

Problem: If you use a non-existent macro in a code module, the Map Designer cannot reopen that map.

Solution: Checking for file existence of a string composed purely of a non-existent macro caused an infinite loop. Fixed the loop and ensured that an empty string cannot be seen as an existing filename.Date

Fixed: 2005-02-01

Problem: Map with IBM DB2 Universal Database Connector as target fails to run. It displays a Segmentation Fault error.

Solution: The problem is that a string consisting of about 120 digits was being passed to CInt. The copy routine uses a buffer of 101 characters, allowing a maximum number of 100 digits. However, two functions were calling the copy routine passing only a 100-character buffer. Thus, the trailing null was overflowing the buffer. Changed the buffers being passed to 128 bytes.Date

Fixed: 2005-02-10

Problem: When changing a line style from "pattern" to "pattern and relative position," an error occurs and the changes are lost.

Solution: The selected column spontaneously shifts to the first column and makes the same changes, which results in the error. The code has been changed, to save the selected column and make sure it is set correctly, before making the changes.

Date Fixed: 2005-02-0

Problem: When a regular expression is used as the start rule of a field and the Include Tag selected, the data is not extracted correctly.

Solution: The regular expression length, which was being tested to determine the length of the start tag, caused the problem rather than the length of the string matched by the regular expression. Fixed code to address this problem.

Date Fixed: 2005-02-2

Problem: When connecting to a log file using the "Data Junction Log File" connector in Map Designer, the source log file is not parsed correctly when viewed in the source data browser.In addition, if you click on the Structured Schema ellipsis, the Extract Schema Designer is displayed and there are no Accept record actions on the left side of the screen. Also, after the third record, some of the key fields are blank.

Solution: Created a new log file extraction script and added new functionality. It does everything the old script did, plus has two extra record types to collect product information like Owner type of Engine and Version, and to collect Process information like when a scripting step started, or what a decision step evaluated to.

Date Fixed: 2005-02-1

Problem: Data is not being cleared between repeating loops in the map.

Solution: Changed code to address this problem.

Date Fixed: 2005-02-0

Problem: If Visual SourceSafe is not configured correctly, the Repository Explorer crashes.

Solution: The Visual SourceSafe interface in Repository Explorer has been completely re-designed.

Date Fixed: 2005-02-2

Problem: TimeValue function does not work.

Solution: The function did not work because there was no space before the PM ("10:26PM"). Made changes to the code.

Date Fixed: 2005-02-1

Problem: Format function using exponents "e+" does not work correctly.

Solution: Fixed code to format scientific notations correctly.

Date Fixed: 2005-02-2

Problem: A process with an Oracle connector with an invalid database name displays 'Segmentation fault' error.

Solution: Changed code to fix this problem. The process now displays the error "Error opening data Source." Return Code: 25519.

Date Fixed: 2005-02-0

Problem: Either the message objects or buffers are not being cleared.

Solution: Fixed a problem with decimal fields that were not being cleared.

Date Fixed: 2005-02-1

Problem: Transformations that use a query file for the source will not run in a process.

Solution: Fixed a problem where transformations that use a source query file are prevented from executing.

Date Fixed: 2005-02-0

Problem: If you store a message object in a process, try destroying and then running it again. The message persists, and does not get overwritten.

Solution: Modified the FTP connector to clear the FTP cache on disconnect.

Date Fixed: 2005-02-0

Problem: "RETURN_CODE_25504" step returns value of "False" and terminates without completing rest of process.

Solution: Fixed a problem tracing execution paths when a step has IgnoreError set.

Date Fixed: 2005-02-0

Problem: Install shows wrong minor version number.

Solution: Updated the package script for the new version numbering scheme.

Date Fixed: 2005-02-2

Problem: The bar graphs on the Iteration and Time histogram tabs do not display correctly if the scrollbar exists at the bottom of display.

Solution: The minimum box width in the histogram is set to six pixels, ensuring that at least four pixels of color will appear.

Date Fixed: 2005-02-1

Problem: The SWIFT source data is displayed incorrectly in the browser after connecting to it via the SWIFT.2003.MT202.0.0.ds.xml schema.

Solution: The problem existed with the SWIFT connector (not SWIFT schema). Changed code to fix this problem.

Date Fixed: 2005-02-1

Problem: Help page is not displayed for Find/Data in Source View Designer.

Solution: The broken link has been fixed. The correct Help page is now displayed.

Date Fixed: 2005-03-0

Problem: Pressing help button in Process Designer fails to display the help page.

Solution: The broken link has been fixed. The correct Help page is now displayed.

Date Fixed: 2005-03-0

Problem: Broken help link for "Use an External Data File Schema".

Solution: The broken link has been fixed. The correct Help page is now displayed.

Date Fixed: 2005-03-0

Problem: When selected Help > Contents from the toolbar, the help page is not found. This is also true with respect to "?" in the toolbar menu.

Solution: The broken link has been fixed. The correct Help page is now displayed.

Date Fixed: 2005-03-0

Problem: The "Validate" window in Join Designer has a broken help link.

Solution: The broken link has been fixed. The correct Help page is now displayed.

Date Fixed: 2005-03-0

Problem: In Map designer, Source Connection tab, select "Copy Schema..." button. Click the "Help" button on the bottom of the page. The Help page is not displayed.

Solution: The broken link has been fixed. The correct Help page is now displayed.

Date Fixed: 2005-03-0

Problem: When a RIFL function is invoked and Help button is selected from the function, the Help page is not displayed for the Date/Time category.

Solution: The broken link has been fixed. The correct Help page is now displayed.

Date Fixed: 2005-03-0

Problem: When a RIFL function is invoked and Help button is selected from the function, the Help page is not displayed for the Math category.

Solution: The broken link has been fixed. The correct Help page is now displayed.

Date Fixed: 2005-03-0

Problem: When a RIFL function is invoked and Help button is selected from the function, the Help page is not displayed for the Parse category.

Solution: The broken link has been fixed. The correct Help page is now displayed.

Date Fixed: 2005-03-0

Problem: The help file is not found in the Upgrade Utility.

Solution: The broken link has been fixed. The correct Help page is now displayed.

Date Fixed: 2005-03-0

Problem: Open Map, select Data Browser and click on Help. The help page is not displayed.

Solution: The broken link has been fixed. The correct Help page is now displayed.

Date Fixed: 2005-02-2