Release notes for portingset: 7.3a.02/SUN_SPARC/SOLARIS2.6/PA.2147Release notes for portingset: 7.3a.02 MaintReger: # 16075 (BDUX9547): Repair -bse argument parsing in bshell. Date: Fri, 22 Nov 2002 13:29:46 +0100 Created on: DevEscher Type: bugfix Problem Description (Customer terms) Calling bshell with the -bse ... doesn't work. Problem Description (Technical terms) The bshell should handle the -bse argument in a very early state during startup. A previous fix in the bshell (BDUX9290) caused however that the logging module of the bshell was being initialised before handling the -bse argument. This caused that setting the BSE environment with the -bse argument didn't work anymore. Test Procedure Start bshell with the -bse argument. MaintReger: # 15878 (BDUX9366): Oracle installer fails on duplicate temporary files Date: Wed, 6 Nov 2002 17:30:37 +0100 Created on: DevEscher Type: bugfix Problem Description (Customer terms) Oracle error ORA-1537 during Baan installation on Oracle 9 when a TEMPORARY file is added that already exists in the database. Lower Oracle versions and other file types do not cause an error. Problem Description (Technical terms) In Oracle 9 a new oracle view has been introduced for temporary data files. The existing data files view does not show temporary files anymore. Workaround Be sure you add non existing files. Test Procedure Run the installer against Oracle9, and add a file to the TEMPORARY tablespace that already exists. The current installer fails with ORA-1537 (dbs_errno 2537). The new installer succeeds. Motive source TCS:100-128805 MaintReger: # 15868 (BDUX9364): Repair timetest.c Date: Wed, 6 Nov 2002 10:01:46 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) The time interval change for AS/400 (in port_BDUX9333) contained an syntax error. Test Procedure clearmake libs MaintReger: # 15863 (BDUX9327): ASM server host<->ip detection fix Date: Tue, 5 Nov 2002 16:39:17 +0100 Created on: MaintReger Type: bugfix Problem Description (Customer terms) When you use an IP Adress instead of a system name in the MMC-Snapin of ASM, it will fail. ASM is not able to connect and will not start. In the MMC snap-in you will see a read-cross. Problem Description (Technical terms) The CompareHostName() function compares only host strings and doesn't care about IP addresses. Workaround Convert the incoming host names into IP addresses and compare them. Test Procedure On the ASM snapin MMC console use az IP address for the host system. Motive source DF:128446 MaintReger: # 15855 (BDUX9354): Update of timezone information Date: Tue, 5 Nov 2002 15:30:23 +0100 Created on: MaintReger Type: bugfix Problem Description (Customer terms) The summertime in Brazil is defined to be at November 3rd, 2003 ( at midnight ). In the $BSE/lib/zonerules/southamerica file, the last entry found is for year 2000, First Sunday in October. Problem Description (Technical terms) Download file ftp://elsie.nci.nih.gov/pub/tzdata2002d.tar.gz and unpacked it. Copied new timezone rule files to /vobs/logic/tz/zoneinfo/zonerules Compiled new zoneinfo files Add new entries to deliver.lst Workaround Not applicable Test Procedure Not applicable Motive source DF:210812 MaintReger: # 15853 (BDUX9294): Fixed problems in BDB libraries Date: Tue, 5 Nov 2002 15:22:29 +0100 Created on: MaintReger Type: bugfix Problem Description (Customer terms) 1) On Windows2000, the library libsql_c is missing from the BDB API libraries. 2) The following functions are missing in shdbapi.dll: baan_timezone_set double_out float_out TSSsetlocale TSSdontreadunicode close_stream_servers num_to_date al_utc_to_local date_to_num al_unpack_utc al_pack_utc baan_encrypt TSStssmbs_n2extmbs_n 3) The USER_INFO structure cannot be accessed when user the shdbapi.dll and the following information from this structure is required: language locale package combination timezone user name user type MaintReger: # 15849 (BDUX9333): Add .bnddir file to savf file (AS/400) Date: Tue, 5 Nov 2002 11:55:29 +0100 Created on: MaintCorelli Type: porting specific change Problem Description (Customer terms) Vertex API can not be created for AS/400 becase of missing binding directories for libalbsh, libhal, libtss, libipc, libtz, libyybsh and libtvtx libraries Problem Description (Technical terms) .bnddir files were not included in the stream files containing the above mentioned libraries. Workaround In AS/400 portingsets 7.1c.05 and 7.1d the binding directories are missing from every save file resotred from API.TAR. These binding directories must be created by hand. Use the following instructions to do so: 1. The name of the binding direcotry have to be the same as the save file. e.g. if the save file called LIBALBSH, then the binding directory must be called LIBALBSH. 2. The list of modules to be added to the binding directory can be obtained from the save file itself. Use command DSPSAVF for this. 3. If you have the list of objects, then first you must create the binding directory with CL command CRTBNDDIR (e.g.: CRTBNDDIR REL7.1D/LIBALBSH) 4. Finally all the modules have to be added to the binding directory with CL command ADDBNDDIRE. (e.g. ADDBNDDIRE BNDDIR(REL7.1D/LIBALBSH) OBJ((REL7.1D/AL_AMEM6A *MODULE)) 5. The creation of the binding directory and adding the modules have to be done for every restored save file you need to bind. Because of the large number of modules to be included per binding directory, it is advisabe call the commands from withing QSH with the help of Unix like utilities (see system command). E.g. command 'system DSPSAVF REL7.1D/LIBALBSH | grep \*MODULE' lists all the modules from save file LIBALBSH in library REL7.1D. It is much easier to work with the output of the CL commands in QSH." Test Procedure Create a new poringset .tar file (/home/port/bdux9333/porting/OS4004.5_PA.XXXX.tar) Create a temporary library XXX pax -v -r -C 819 -f OS4004.5_PA.XXXX.tar API.TAR pax -v -r -C 819 -f API.TAR api/lib/libalbsh.a RSTFRMSTMF RSTLIB(XXX) STMF('/home/port/bdux9333/porting/api/lib/libalbsh.a') DSPSAVF FILE(XXX/LIBALBSH) Motive source DF:210516 MaintReger: # 15844 (BDUX9330): Adding iget/sget.field.attr + additional form.control trace Date: Tue, 5 Nov 2002 10:26:14 +0100 Created on: MaintReger Type: bugfix Problem Description (Customer terms) 1) in the 4GLE a lot of field loops are done while testing only one or two field attributes. This means a lot of select.uniq.field calls with result in the copying of many attr.* variables. 2) impossible to know from -dbgfun/-dbgcpu which of the 41 form.control functions are used Problem Description (Technical terms) 1) performance killer. 2) the form.control types are in IDX(1). MaintReger: # 15842 (BDUX9342): Merge from project bugf_BDUX9334, Again licd6.x core dump Date: Tue, 5 Nov 2002 10:24:01 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) License deamon licd6.x can dump core when running in a cluster and connections get lost. Problem Description (Technical terms) The connection with the oserver could be lost when waiting for messages from the oserver. Workaround Do not use licd6.x in cluster. Motive source DF:938501 MaintReger: # 15815 (BDUX9315): Leave content file in place Date: Wed, 30 Oct 2002 13:37:35 +0100 Created on: DevEscher Type: porting specific change Problem Description (Technical terms) Leave content file containing what's delivered in place. This helps for later checks.. MaintReger: # 15806 (BDNT9301): licd6.1, ignore SIGPIPE and extra logging bis Date: Wed, 30 Oct 2002 09:57:04 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) licd6.2 did not compile on NT because of unresolved symbols Workaround Use #ifdef __WIN32__ at the corrct place Test Procedure Build of licd Motive source DF:938501 MaintReger: # 15798 (BDUX9246): repgen fails if a underscore character is used in input field name Date: Tue, 29 Oct 2002 14:23:50 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) If you define a report with a before.field layout and the before.field layout is for a field containing an underscore then the sections in a report script are not taken for this layout and the repgen stopps with the error No current layout. Problem Description (Technical terms) Only alphanumeric characters and the "." character are allowed in sections, currently. Workaround If you need to write statements in your reportscript you should define report input fields without underscore characters in his name. Motive source DF:129453 MaintReger: # 15788 (BDNT9286): Fixed bug in bshell process affinity list. Do not write entries with socket name in it in the registry, causing a full registry Date: Tue, 29 Oct 2002 11:01:04 +0100 Created on: DevCorelli Type: bugfix Problem Description (Customer terms) Fixed bug introduced by project BDUX7093. Wrong values were written in Windows registry for processor affinity. Each process got a new value, which might cause a registry which runs out of space in worst case. Problem Description (Technical terms) see above. Test Procedure Run a bshell on a fat client and check that no keys in the form are created under HKLM\Software\Baan\\ProcessAffinity MaintReger: # 15784 (BDUX8929): licd, ignore SIGPIPE and extra logging (merge from bugf_BDUX8142) Date: Tue, 29 Oct 2002 10:04:15 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) In a cluster of two and more cooperating licd6.1 processes on Digitals OSF1, the licd6.1 processes could go exit unexpectedly. Also they could loose track of each other. Problem Description (Technical terms) A SIGPIPE could occur when a licd6.1 writes to a socket that is already closed by the other side. This signal was not handled so the process would go exit without logging any reason. Communication between licd6.1 processes is also a major problem. Deadllocks between servers can easely occur. The recovery lead to losing contact between servers. Workaround Signal handler for SIGPIPE, watermark mechanisme to decrease communication between license servers. Test Procedure Run licd6.2 in cluster on Tru64 Motive source DF:938501 MaintReger: # 15779 (BDUX9290): Repair argv[0] setting (merge from bugf_BDUX9285) Date: Tue, 29 Oct 2002 09:25:47 +0100 Created on: MaintReger Type: bugfix Problem Description (Customer terms) Stack traces of the bshell on HP were no longer logged in ${BSE}/log/log.bshell. Problem Description (Technical terms) On HP the call U_STACK_TRACE() did no longer work because the argv[0] was mangled. Workaround Do not change argv[0]. MaintReger: # 15726 (HYUX9220): Removal of Statement Cache Functionality Date: Wed, 23 Oct 2002 14:37:00 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Core dump during import process. Problem Description (Technical terms) Statement cache functionality was not properly implemented to handle all the cases and because of this driver dumps a core sometimes. Workaround Disable the statement cache functionality by including the following in db_resource file. inf_use_prepare_cache:0 Test Procedure Enable the statement cache functionality and try running a query. Check the dbslog file and ensure statement cache functionality is ignored. Motive source TCS:100-128654 MaintReger: # 15721 (BDUX9235): Add log message when text resources are not supported for a platform Date: Wed, 23 Oct 2002 12:47:08 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Added log message if text resources from $BSE/lib/textinf are not supported on a platform Problem Description (Technical terms) see above Test Procedure see defect text in scopus Motive source DF:130037 MaintReger: # 15716 (BDUX9232): HOME and USER are not correctly set when using baan login daemon Date: Wed, 23 Oct 2002 12:34:05 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Environment variables HOME and SHELL where not correctly set, when logging in with Baan Logon Daemon on a Unix system (e.g. AIX/HP). Or the variable(s) are sometimes set. Problem Description (Technical terms) Local variable was passed to putenv(), causing unreferenced/empty environment variables in the system environmet. Workaround use rexec protocol. Motive source DF:129988 MaintReger: # 15691 (BDUX9205): bshell hangs when it can not start JVM Date: Tue, 22 Oct 2002 12:19:24 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) Wnen the correct java version was not available or could not be started, the beshell would hang or core dump. Problem Description (Technical terms) Cleaning up the Java structures after a failed JVM startup would lead to destroying a semophore twice which lead to memory alocation errors. Workaround Do not call destroyJavaQueue(execJavaRequestQueue) after stopJavaVM. Test Procedure Start bshell and try to run tmboaserver with an incorrect jvm_options file. Motive source TCS:100-128656 MaintReger: # 15657 (BDUX9163): AS400, enabling java 1.2 for V5R1 Date: Mon, 14 Oct 2002 17:14:10 +0200 Created on: MaintReger Type: porting specific change Problem Description (Customer terms) bshell6.2 only runs with java version 1.1.8 when compiled on V5R1 Problem Description (Technical terms) Configuration file ibm_as400V5R1.cf does not contain the -DOS400_JVM_12 flag in the ExtraJavaDefines define. Workaround Use Java 1.1.8 Test Procedure Build Motive source DF:208174 MaintReger: # 15653 (BDUX9155): NQP: remove hash statistics Date: Mon, 14 Oct 2002 15:13:30 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) Statistics were kept in memory about the usage of hash structures. These statistics were never used, but consumed a considerable amount of memory (a simple order entry session as used in the "Reger Performance" project showed that about 200K byte was used for these statistics). Test Procedure Compile a bshell/driver both with and without the HASH_STATS_ON flag. Run a simple session (like order entry) and observe an increase in memory usage when using the bshell/driver compiled with HASH_STATS_ON flag. MaintReger: # 15651 (BDUX9138): Enable java 1.2 for AS/400 Date: Mon, 14 Oct 2002 15:06:59 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Customer terms) On AS/400 default Java version 1.1.8 was used when running the bshell java integration. Problem Description (Technical terms) On AS/400 the java interpreter is a subsystem that can run multiple java versions. Default that is the highest 1.1.X version. To override the version you can start the java virtual machine (jvm) with the -Djava.version=1.2 option. Workaround Compile with the JNI_VERION_12 define and use option -Djava.version=1.2.2 when starting java subsystem. Test Procedure Create a bse environment to run the bshell. Motive source DF:208174 MaintReger: # 15632 (BDUX9100): SGI java home Date: Thu, 10 Oct 2002 17:35:13 +0200 Created on: MaintReger Type: porting specific change Problem Description (Technical terms) compilation error BCLM Test Procedure compile MaintReger: # 15625 (BDNT9129): improve Delivering on NT systems Date: Thu, 10 Oct 2002 10:37:15 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) Cameo examples doesn't compile. Problem Description (Technical terms) Modified include files occour this problem. Workaround Fix a typo in delivery.lst, add one header more to the delivered headers and modify deliver.exe for compile the delivered examples. Make deliver.exe capable to compress the delivered data into one tar.Z file. Test Procedure Do a nightly build with the new deliver.exe and use the new functions which are described in sample.cfg. MaintReger: # 15615 (BDUX9131): Fixed compilation errors on Unix after merge of BDNT9029 Date: Wed, 9 Oct 2002 15:21:24 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Fixed compilation errors on Unix Problem Description (Technical terms) see above Test Procedure Compile audit stuff (db/server/AUDIT) MaintReger: # 15607 (BDNT9121): Fixed bug in SHM caused by project BDNT9029 Date: Wed, 9 Oct 2002 11:00:28 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Bug in creating mutex for Shared memory. Using wrong name. Problem Description (Technical terms) see above. The slash in word 'Global\' should not be converted on Windows 2000, since it has a specicial meaning for Windows when creating mutex (in combination with terminal services). Test Procedure Very difficult to test. If using terminal services, the SHM mutex is now created in the global namespace. This is OS dependent, for W2k + terminal services. MaintReger: # 15602 (BDUX9077): enable report language for UTC time and date Date: Tue, 8 Oct 2002 16:10:58 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) The maintain language session is set up with different seperators for germany language. In maintain date formats I choose language separator ( Help says that it then will take the user!! language, which always is English, but again in Baan 4 it takes the report language). We convert to runtime DD, but the formats are English formats all over. Problem Description (Technical terms) When in the format both the UTC date and UTC time is used the language of the report is not taken. when this is forced the language is only taken for the time, and not for the date that is displayed first. format can be seen as %u001 %U001 but the language option is added as: %u001 %U001,1 Workaround Specify the date as last option of a field ( make 2 fields, one for time and one for date ) and force a language option in the session. Test Procedure compile a report with format %u001 %U002 in an other language of the user, start this report also in the other language and see if the date is displayed in the correct language. also check the output of the next line: function main() { string rjr(32) rjr = sprintf$("%u001,1", 1033040301) rjr = sprintf$("%U002,1", 1033040301) rjr = sprintf$("%u001 %U002,1", 1033040301, 1033040301) } execute this as a user in an other language the language 1 and see if the correct language seperators are used for all 3 lines. Motive source DF:129085 MaintReger: # 15584 (BDNT9028): XML: test string correctness early Date: Tue, 8 Oct 2002 10:50:04 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) need a debug flag to verify strings at the moment that they are added to an XML tree, instead of when the tree is serialized Test Procedure use -dbgxml flag MaintReger: # 15582 (BDNT9029): Fixed problems in release builder. Shm failed to create mutex. Date: Tue, 8 Oct 2002 10:04:56 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Release builder didn't work correctly. Problem Description (Technical terms) The release builder was not able to install a Baan environment, because shared memory didn't start up. After days of research, this was caused by a wrong dump of the Unix tape (all binary files were corrupted, probably caused by ASCII ftp). However, SHM stopped without any error, therefore enhanced logging. During research, some other bugs regarding logging were discoverd (e.g. when a u file contained wrong data). All problems where Windows NT related (problems do not occur on Unix). Workaround Put the right tape dump there, and no error will occurr. Test Procedure actually no bug was fixed, only logging. You can test this: 1. put a wrong tss_c6.2 file in $BSE/lib. SHM should now exit whith a clear log message in the Event viewer. 2. remove an entry e.g. pacc: in u file in $BSE/lib. A clear log message should be put to Event viewer. 3. Run release builder: dialogs should not be but on top of other dialogs (progress indicators). No broken texts in dialogs anymore. 4. Although not fully confirmed: made creation of Mutexes for SHM more Windows 2000 (Terminal Server) compliant. Add Global\\ directive. See also MSDN on CreateMutex() function. MaintReger: # 15558 (BDNT9066): Textwidth not calculated correctly because of vertical scrollbar in multiline text window Date: Fri, 4 Oct 2002 08:17:04 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) When in a multiline textwindow the maximum number of columns is determined to 80 and no vertical scrollbar is present the number of characters a user can type is 82. After savng the text and reopening it all lines are cut off at collumn 80, possibly resulting in a wrong layout. To prevent this the vertical scrollbar should always be present, but disabled when the number of lines exceeds the maximum number of lines per page, then the scrollbar needs to be enabled. Problem Description (Technical terms) The vertical scrollbar in a multiline textwindow is not controllable. Windows removes the scrollbar when the complete text fits in the window. When this is not the case the scrollbar is made visible. But when the scrollbar is not visible the user can type over the space where the scrollbar should be. Workaround Not applicable Test Procedure Run Baan (Corelli or higher) from any server using BW and start session tttxt1100m000. Search or create a text with textgroup 'text' (meaning the text is editted using the Baan editor). Choose 'Edit Text' form the specific menu. Look at the textwindow and check if there is a vertical scrollbar. If not, you can type more than the defined number of characters on one line. If the number of lines exceeds the maximum number of lines per page a vertical scrollbar is made visible. Characters positioned on the last two columns of the textwindow are moved to the next line. Motive source DF:127147 MaintReger: # 15548 (BDUX9071): compilation error xml Date: Thu, 3 Oct 2002 13:35:15 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) libxml does not compile on SGI, which is needed for BCLM build. encoding is a member of a struct and also a define Test Procedure compile MaintReger: # 15532 (BDNT8979): Core dump during exchange import/export as normal user Date: Tue, 1 Oct 2002 16:15:58 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Normal users trying to Import/Export data with exchange lose their connection to the Baan environment (session and Menu Browser disconnect). Problem Description (Technical terms) During export/import data with exchange the function exec_dll_function() is used. The first argument points to the DLL that hosts the function which has to be executed. During the permission check to this DLL a memory overflow can appear in case of DLL name (incl. path) is longer as 22 characters. Workaround n/a Test Procedure 1.) Compile following program (bic test.src -o testo) Source of test.src: function main() { long ret ret = exec_dll_function("C:\tmp\extra_long_path_for_exchange_core_file_test\mydll","func_in_dll") if ret = 0 then message ("Okay") endif } 2.) Create the directory used in previous exec_dll_function() call. If you are using another directory ensure that the length of the directory name is longer as 22 characters. 3.) Compile following DLL (bic mydll.src -o mydll) Source of mydll.src : function extern func_in_dll() { } 4.) Ensure that you are a normal user with permission to do everything. 5.) Run program testo from GUI. The message "Okay" should be displayed. Motive source DF:210062 MaintReger: # 15527 (BDUX9023): Fix makefile for taxware shared library build Date: Tue, 1 Oct 2002 10:45:33 +0200 Created on: MaintBaanIVc Type: bugfix Problem Description (Technical terms) Due to how Taxware compiles their objects, some of their objects will not work with other objects. They do, however, deliver versions of these objects that work together in another directory. Workaround The makefile can be altered by the user if they know where they need to look for the right objects. Test Procedure Make the required changes to the makefile to point at the correct base directories and then run make to confirm that the api compiles. MaintReger: # 15513 (BDNT9033): bic: repair error message about standard.secu Date: Fri, 27 Sep 2002 13:53:05 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) bic complains: Fatal Bic Product Id Security Error: invalid file '$BSE/sec/standard.secu'. The mentioned file name is incorrect; it should be '$BSE/secu/standard.secu'. Problem Description (Technical terms) Error message has incorrect text. Test Procedure change one byte of the file '$BSE/secu/standard.secu', and run bic for some session to generate the error message MaintReger: # 15512 (BDUX9032): Introduce where-used for labels. Date: Fri, 27 Sep 2002 12:04:01 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) >From Reger onwards, labels can be used in Program scripts. The usage of these labels could not be made visible in the where-used output. Problem Description (Technical terms) This is a manual backport of the corresponding Gemini project enha_BDUX9020. Test Procedure Add some pragma's and whereused constructions to your program script and compile with the -u and -f flags. Check the where-used output. Motive source TCS:106-127039 MaintReger: # 15481 (BDNT8989): Add CleanUp functionality for Delivering Date: Wed, 25 Sep 2002 08:34:42 +0200 Created on: MaintBaanIVc Type: bugfix Problem Description (Technical terms) If a component of a porting set have been removed then in a nightly buid doesn't remove that component from the target directory on NT systems. Workaround Making a new cleanup functionality for erasing the destination path before delivering. Test Procedure set the CLEANUP flag TRUE in the .cfg file. MaintReger: # 15478 (BDUX8926): reconfig does not always flush dbs_dd cache Date: Tue, 24 Sep 2002 14:59:34 +0200 Created on: DevEscher Type: bugfix Problem Description (Customer terms) Some tables that are reconfigured are not invalidated in the shared memory cache of the database drivers resulting in error E_BDB_DD_CORRUPT (error 512). Problem Description (Technical terms) When bdbreconfig determines that no changes in the database are needed (like reference changes), no new DD is sent to the driver. As the sending of the new DD triggers the invalidation of the cache, this invalidation will not be done. Workaround restart the shared memory after reconfigurations that result in E_BDB_DD_CORRUPT (error 512). Test Procedure Copy a dd to dd.new: eg ttadv200.new Add to $BSE/lib/compnr6.2: ttadv200:300:100 run 1: DBSLOG=1000 BDB_DEBUG=02 USE_SHM_INFO=1 bdbreconfig6.2 -n -N ttadv200 -C 000 002 005 100 The output should contain lines like: -> EXE_SYSTEM on server 400ffe80 code 8 msglen 12 (ttadv200.000) <- EXE_SYSTEM answer len 0 () -> EXE_SYSTEM on server 400ffe80 code 8 msglen 12 (ttadv200.002) <- EXE_SYSTEM answer len 0 () -> EXE_SYSTEM on server 400ffe80 code 8 msglen 12 (ttadv200.005) <- EXE_SYSTEM answer len 0 () -> EXE_SYSTEM on server 400ffe80 code 8 msglen 12 (ttadv200.100) <- EXE_SYSTEM answer len 0 () dbs.log should contain lines like: Msg_type 7 received. Invalidate DBS_DD in shared memory. tname = 'ttadv200', compnr = 0 dbs_dd_env = '. . .', dbs_dd 'ttadv200000' Msg_type 7 received. Invalidate DBS_DD in shared memory. tname = 'ttadv200', compnr = 2 dbs_dd_env = '. . .', dbs_dd 'ttadv200002' Msg_type 7 received. Invalidate DBS_DD in shared memory. tname = 'ttadv200', compnr = 5 dbs_dd_env = '. . .', dbs_dd 'ttadv200005' Msg_type 7 received. Invalidate DBS_DD in shared memory. tname = 'ttadv200', compnr = 100 dbs_dd_env = '. . .', dbs_dd 'ttadv200300' Note that the physical company (300) is invalidated instead of the logical company 100. run 2: DBSLOG=1000 BDB_DEBUG=02 USE_SHM_INFO=1 bdbreconfig6.2 -n -N ttadv200 -C 100-120 The output and dbs.log should contain invalidates for all companies in the specified range. Motive source mail about problems with Gemini shuttle 3 MaintReger: # 15447 (BDUX8708): Make recipients for all releases identical Date: Fri, 20 Sep 2002 14:12:10 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) Make receipients for all releases identical for a notification that a build has been done Test Procedure Run sendreport.bat (on NT) or check the mail commands after a portingset delivery has been done. MaintReger: # 15439 (BDUX8971): Audit must write client types into the audit trail Date: Fri, 20 Sep 2002 13:09:33 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) The 3/4 GL code that handles the audit data breaks on columns containing multi-byte strings. Problem Description (Technical terms) The problem is caused by the fact that the audit trail contains a unexpected type as the data type for multi-byte strings. This is an internal 'dbs' type, which doesn't match the 'db' type defines as known by applications. Consequently, Exchange interprets 'db.multibyte' fields as a 'db.text' numeric field. Note: this problem also occurs in previous releases. (It seems Exchange based on audit was never tested in a multibyte environment!) The following was decided: The problem will not be solved in Reger SP1, but it will be solved in the Reger SP2 audit server. The database team will make a proposal for this. This proposal will include the use of the db.multibyte type in the audit trail, but it may include a solution for other db types as well. Consequently, a scope change was requested for the iBaan ERP 5.2a SP1 release. (See PowerDocs #19717). The scope change was officially approved by the ERP PSG on 23 May, 2002. Test Procedure Exchange regression test Motive source TCS 106-124824 MaintReger: # 15426 (BDNT8958): Do not reconnect BW in case of Automation Date: Thu, 19 Sep 2002 12:39:34 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) When Worktop is used and the connection with the bshell aborts (for instance because the Bshell crashes), BW will give the user a chance to reconnect. However the Worktop will also beform some retries. Therefore BW should not reconnect in case the Worktop is used. Problem Description (Technical terms) The fact that BW was started by the Worktop is determined by the fact that BW was started through Automation. In this case BW will only give a message to tell the user that the connection with the Bshell was broken, however the user will not have a chance to reconnect. Workaround Press "NO" when the question comes to reconnect. Test Procedure See problem description above. MaintReger: # 15414 (BDUX8951): Remove Kermit from AS400 build Date: Wed, 18 Sep 2002 13:44:59 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) Build hangs on AS400 systems Problem Description (Technical terms) Error cames during the kermit builds on AS400 systems. Workaround Checking for AS400 build and no make happens on that kind of builds, because it doesn't needed for AS400. Test Procedure Build a portingset for AS400 MaintReger: # 15412 (BDUX8824): Relnotes update Date: Wed, 18 Sep 2002 13:43:16 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) Confusing error messages, because of wrong search by Relnotes script. Workaround Ignore them MaintReger: # 15405 (BDUX8946): build error on old sun box Date: Tue, 17 Sep 2002 13:41:59 +0200 Created on: MaintBaanIVc Type: porting specific change Problem Description (Technical terms) Compiler with -xtarget=ultra flag produces binaries which do not run on our old build box 'sun'. Workaround Buy a new build box. MaintReger: # 15371 (BDUX8863): wrong LD_LIBRARY_PATH Date: Wed, 11 Sep 2002 15:58:47 +0200 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) bshell crashes when tmboaserver is started Problem Description (Technical terms) When LD_LIBRARY_PATH is composed by bshell, a bufferoverflow can occur. ipc_boot does not put LD_LIBRARY_PATH in environment Workaround Put next line in $BSE/lib/bse_vars: LD_LIBRARY_PATH=/bse_dir_of_customer/shlib:/usr/baan/shared/shlib:/usr/java/lib or LD_LIBRARY_PATH=/bse_dir_of_customer/shlib:/usr/baan/shared/shlib:/usr/j2se/jre/lib/sparc:/usr/j2se/jre/lib/sparc/client:/usr/j2se/jre/lib/sparc/native_threads Test Procedure start tmboaserver on Solaris with JRE1.3 MaintReger: # 15365 (BDNT8895): Fix function call convention for ARM shlib: affects only NT Date: Wed, 11 Sep 2002 15:03:38 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) 1: Lost connection with BW when ARM is enabled on Windows. 2: Ntbshell can not find/open arm.dll. Should look for libarm32.dll Problem Description (Technical terms) 1: On Windows function call in DLLs use other call convension than ordinary C. The proto type of function calls in arm.h was OK, but proto type on function pointers to bind calls not. 2: Name of arm lib in art_fun.c was only good for unix. On NT other name should have been used Workaround 1: not possible 2: rename LIBARM32.DLL to ARM.DLL Test Procedure Install Corelli on NT. Enable ARM in $BSE/lib/defaults/all (art_enable:1 art_trace:3) Install ARM software Notice ARM DLL not found Rename ARM DLL to ARM.DLL Notice ntbshell will break. MaintReger: # 15356 (BDUX8892): Update generated _aix files after changes in bic. Date: Wed, 11 Sep 2002 09:50:38 +0200 Created on: MaintReger Type: porting specific change Problem Description (Technical terms) After changing bic.y in project bugf_BDUX8852, the generated files for _aix had to be updated as well. Test Procedure None MaintReger: # 15349 (BDUX8852): Add keyword __OBJECT__ to bic. Date: Tue, 10 Sep 2002 16:55:32 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) There is no way to determine the current object name from withing Baan 3GL. Having this name available would give some benefits in the tools (e.g. in better assert() and in dal.file.to.class.name()). Problem Description (Technical terms) See above Test Procedure Create a small 3GL program that outputs the stringvalue of __OBJECT__. MaintReger: # 15345 (BDUX8889): Fixed error in Makefile Date: Tue, 10 Sep 2002 14:36:11 +0200 Created on: DevEscher Type: porting specific change Problem Description (Technical terms) Fixed compilation problem on non-ClearCase platforms Test Procedure run make -n in porting dir MaintReger: # 15328 (BDUX8554): Removed some old BCBE stuff Date: Mon, 9 Sep 2002 15:27:31 +0200 Created on: MaintReger Type: porting specific change Problem Description (Technical terms) Removed old BCBE stuff (orb bshell, xma bshell, idl compiler) sources Test Procedure check source code that bcbe stuff was not there anymore MaintReger: # 15327 (BDUX8872): Oracle 9.2 Date: Mon, 9 Sep 2002 10:18:57 +0200 Created on: MaintReger Type: porting specific change Problem Description (Customer terms) Oracle 9.2 is handled and reported as Oracle 9.0. Problem Description (Technical terms) The Baan/Oracle drivers using the shared Oracle libraries (7.3x porting set and higher) do not recognize Oracle 9.2 yet. MaintReger: # 15318 (BDUX8845): Cleanup Project.all Date: Thu, 5 Sep 2002 16:39:39 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) cleanup; no siemens no sequent no bshellorb no bshellxma use oracle 8.1.7.2.0 header files MaintReger: # 15316 (BDUX8855): __FUNCTION__ not recognized anymore by bic. Date: Thu, 5 Sep 2002 14:40:11 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) The keyword __FUNCTION__ is not recognized anymore by bic. Problem Description (Technical terms) Line containing __function__ was accidently removed from the table with keywords in an earlier project. Test Procedure Compile a simple 3GL script that prints __FUNCTION__. It should compile now, and output the name of the function in which it is used. MaintReger: # 15306 (BDUX8641): Investigate core dump in bshell Date: Wed, 4 Sep 2002 16:22:08 +0200 Created on: DevEscher Type: bugfix Problem Description (Customer terms) 1. Core dump in bshell when calling dll functions with illegal number of arguments 2. Whith -dbgobj, -dbgcpu, -dbgsym and tracelevel >= 2, a core dump occurs. Problem Description (Technical terms) see above. This problem only occurs in the programming phase (normally) and is existent for years now. Test Procedure Problem 1: test for programmes: write a scripts which calls a dll function with wrong number of arguments. Check that no core-dump occurs. Normal customers do not have the possibility to check this problem. Problem 2: Run bshell with tracelevel >=2 , -dbgsym, -dbgobj, -dbgcpu and check that no coredump will occur (the problem was found while running session: Motive source TCS 106-127008 (medium severity) MaintReger: # 15302 (BDNT8834): Added target to build, since a dependency was removed by ASM code cleanup Date: Wed, 4 Sep 2002 15:08:47 +0200 Created on: DevCorelli Type: bugfix Problem Description (Technical terms) Fixed delivery and build issues on NT. Test Procedure Do a build and check that delivery will succeed without errors/warning MaintReger: # 15296 (BDUX8830): Increase the size for send.message(). Date: Wed, 4 Sep 2002 15:03:23 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) For a new tax integration, the buffer used for send.message() and recv.message() are too small (1K). Must be enlarged to 4K. Problem Description (Technical terms) See above. Test Procedure None MaintReger: # 15292 (BDUX8833): Error handling changes in bdbreconfig Date: Wed, 4 Sep 2002 13:12:11 +0200 Created on: DevBaanIVc Type: bugfix Problem Description (Customer terms) 1. "bdbreconfig -I inputfile" immediately stops after an error occurs. All tables that are not processed yet are not reconfigured. This is not workable for customers, as the reconfiguration process has to be many times to get a single problem at the time. 2. Restarting bdbreconfig might lead to dataloss when a dump file is not removed because the restore action of first run did fail. The second run will overwrite (truncate) the saved dump. 3. When a table does not exist in the RDBMS, the reconfiguration is logged as error (when -i option is not set) Workaround Do not reconfigure everything in one big batch, but run multiple smaller batches. Test Procedure Test procedures of projects: BDUX5851: exit code for bdbreconfig BDUX6247: exit code rcf and -i flag BDUX6691: Bdbreconfig changes MaintReger: # 15283 (BDUX8822): Bic crash problem Date: Tue, 3 Sep 2002 15:55:47 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) bic6.2 dumps core when compiling the following source: #define crash id long function crash void bla(id i) {} Problem Description (Technical terms) In the BICYACC function the an addfunction is called with an empty (0) bic.currtab variable. Workaround Checking for this variable, and if it's empty, give an error message and finish the compiling. Test Procedure Compile the short source in the Problem Description with the bic. Motive source TCS:106-127019 MaintReger: # 15280 (BDNT8811): keystrokes in listbox in Grid Date: Tue, 3 Sep 2002 10:31:57 +0200 Created on: DevEscher Type: bugfix Problem Description (Customer terms) 1. An error occured when using the 'page down' key and 'page up' key in the 'Group Type' Field (dropdown listbox in Grid) this resulted in an assert error. 2. related to this: keyboard support for the listbox in Grid doesn't function correctly. One should be able touse the 'up' & 'down' arrow keys to move through a list (box). Similar to the way it works for the 'find window' and single-occ sessions. Where did the defect occur? (Session code, etcetera): tdpcg115m000 Promotion Groups - when adding a new record Workaround Don't use these keys when focus is on dropdown listbox in Grid. Test Procedure See problem description. Motive source TCS:106-126860 MaintReger: # 15267 (BDUX8671): Multiple printouts after changing paper type Date: Mon, 2 Sep 2002 14:44:27 +0200 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) When someone sends a print job to a device of type printer and the printer is currently down, e.g. the papertype will be changed (ttaad3102m000), sometimes the report will be printed several times if the printer returns to online. Problem Description (Technical terms) In case the pdaemon is busy with converting a print job and the printer returns to online, the print job has to be restarted. (convert process will be killed and job status will be reseted to WAITING in table ttaad320). It was missed to delete the print job from the internal queue. As consequence several entries for the same print job are in the internal queue. The internal queue will be used to determ what has to be printed, so multiple prints are done. Workaround n/a Motive source DF:127542 MaintReger: # 15258 (BDUX8771): buffer overflow Date: Mon, 2 Sep 2002 12:32:03 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) bshells dumps core when SHLIB_PATH, LIBPATH or LD_LIBRARY_PATH contains more than 255 bytes. Problem Description (Technical terms) Lengthcheck of SYSLIB_PATH is wrong, compared with MAX_CLASSPATH. Workaround Make sure your PATH is short Test Procedure Put a very long library path in your environment and start Open World. MaintReger: # 15253 (BDUX8804): Manual merge of project BDUX8785 (LTOE problem in bic). Date: Mon, 2 Sep 2002 10:50:21 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) Bic compiler refuses passing a domain via ltoe() to a ref-argument of a function. This construction was previously allowed. Problem Description (Technical terms) Problem is caused by an earlier fix (BDUX8329). That fix was trying to solve a core dump when compiling the following piece of code using the -u option: function f ( domain tcyesno yn) { } function main () { f( ltoe( 1)) } In the compiler branch for LTOE, the internal variable $$.value.ptr was now initialized to NULL. This works for the above example, but goes wrong if a variable is passed. In that case, $$.value.ptr pointed to the symbol of that variable and is thus lost. When passing the result of ltoe() to e.g. a 'ref'-variable, the compiler will complain that a 'ref' argument is needed. See test-scenario below. Test Procedure Please compile the following example with 'bic6.2 -u -f my_wh -o./tst tst.bc' function negate(ref domain tcyesno n) { n = (n = tcyesno.yes ? tcyesno.no : tcyesno.yes) } function f ( domain tcyesno yn) { } function main () { domain tcyesno flag string a(10) a(1;2) = "hallo" | Make sure the other usage of 'longexp' in | bic still works. f( ltoe( 2)) | Original call: coredump when compiling | this construction with -u option. flag = tcyesno.yes | Resulting call: this construction was refused negate(ltoe(etol(flag)))| Should work now. } Motive source TCS:106-127201 MaintReger: # 15243 (BDUX8793): UTC date problem fix modification Date: Fri, 30 Aug 2002 15:37:12 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) The variable copy was disturbing when you just clicking on it. Problem Description (Technical terms) Any kind of variable select, the variable's name were updated into the history and the command line. Workaround Do this event only when the user clicked with the right mouse button. Test Procedure See the difference, when you select a variable with the left or the right mouse button. Motive source TCS:100-128-285 MaintReger: # 15234 (BDUX8426): Enable multibyte solution in DB2/400 Reger driver. Date: Thu, 29 Aug 2002 14:28:13 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) The Reger DB2/400 driver did not support multibyte fields in table in single byte environment. Problem Description (Technical terms) The TSS_STRING fields were mapped to DBS_STRING type, because the driver did not handle the multibyte table fields (GRAPHIC type on engine level) correctly. The memory allocation for the bind variables were not correct. Also the CCSID should be defined in the sqlda structure before calling the DB APIs. Test Procedure Run the regression test to check the data insert and fetch on multibyte table fields. Motive source TCS:100-127881 MaintReger: # 15216 (BDNT8270): TSS: extend Trad. and Simpl. Chinese Date: Wed, 28 Aug 2002 14:59:16 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) Traditional and Simplified Chinese should be extended. Problem Description (Technical terms) Traditional Chinese (BIG5/CP950) : HKSCS added. Simplified Chinese (GB2312-80/CP936): Microsoft additions of CP936 also added to GB2312-80. Motive source DF:123729 MaintReger: # 15214 (BDNT8776): Added Datacenter certification logo to MMC snapin Date: Wed, 28 Aug 2002 14:52:56 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Customer terms) Added Windows 2000 Datacenter logo in Baan Manager MMC snapin Problem Description (Technical terms) see above Workaround N/A Test Procedure Start Baan MMC manager, and check that a W2k datacenter logo is existent MaintReger: # 15197 (BDUX8759): AS400: Nightbuild fails Date: Mon, 26 Aug 2002 16:50:03 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) Problem Description (Technical terms) The nightly build on AS400 failed. The build detected that SQLparse.y was newer than SQLparse.tab_aix.h, although the last file did not change due to changes in SQLparse.y. Test Procedure Build the SQL library on AS400. MaintReger: # 15179 (BDUX8742): xtarget flag of WorkShop compiler Date: Fri, 23 Aug 2002 14:45:03 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) generated binaries will not execute on buildmachine 'sun' Test Procedure make portingset MaintReger: # 15169 (BDUX8731): SQL: quotes in TSS characters not handled properly Date: Thu, 22 Aug 2002 10:32:09 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) If a query contains a string literal that contains a TSS character, and the TSS character contains a quote (the same character that is used to delimit the string), it is not handled correctly by the SQL processor. Effectively, the quote character is removed from the TSS character. For example, the following string literal contains the TSS character \x9Ba'b. '\x9Ba'bc' When it is used in a SQL statement it results in the following string constant, which contains the TSS character \x9Babc. '\x9Babc' Note: the \x9B character is a single character! Test Procedure Execute the following BSQL script: update dbtst120 set lastname = 'C\x9B!J'B' where empno = 10 ; select lastname from dbtst120 where empno = 10 ; select empno from dbtst120 where lastname = 'C\x9B!J'B' ; The result should be as follows: Processed 1 rows. C\x9B!J'B 10 Note: be sure that the \x9B character is a single character! Motive source DF:119923 MaintReger: # 15167 (BDUX8617): Can not login with blogind when rlogin is disabled on AIX platforms. Date: Thu, 22 Aug 2002 10:28:34 +0200 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) It is not possible to login with blogind when rlogin is disallowed in user security settings on AIX platforms. Problem Description (Technical terms) This behaviour was introduced with BDUX_5612 to check if a account is locked. A additional check for rlogin was build in. This caused the problem. Workaround n/a Motive source DF:127361 MaintReger: # 15123 (BDUX8664): solaris configuration files do not work with /sbin/sh Date: Thu, 15 Aug 2002 11:28:17 +0200 Created on: DevEscher Type: porting specific change Problem Description (Technical terms) Nightly build fails on sun when /sbin/sh is used. Workaround Do not use /sbin/sh or crontab. Test Procedure Nightlybuild MaintReger: # 15098 (BDUX8658): Fix for avoiding qsh: 001-0079 Process ended by exception 9901 Date: Wed, 14 Aug 2002 10:08:00 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) The compiler front-end acc used for native buid on AS/400 did not function correctly. Test Procedure On frite "make World" in rel7.3 and rel7.4a view. MaintReger: # 15093 (BDUX8337): to correct TCS call 106-125976 Date: Tue, 13 Aug 2002 17:29:47 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) The bic_info gave misleading message in case the object was not in proper location. Workaround Checking for the condition what leads for the misleading message. Test Procedure Testing bic_info with giving correct and incorrect object file names as command-line parameters with whole path also in proper and wrong locations Motive source TCS:106-125976 MaintReger: # 15079 (BDUX8639): Can not retrieve java-functions on HP-UX Date: Mon, 12 Aug 2002 15:33:52 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) On HP-UX customer gets message in log file: Load of symbol failed 'JNI_GetDefaultJavaVMInitArgs' Problem Description (Technical terms) libjava.sl is dependant of libjvm.sl, when libjava.sl is loaded, libjvm.sl is also loaded. We can not retrieve libjvm symbols through the libjava handle. Test Procedure tmboaserver MaintReger: # 15053 (BDUX8559): Oracle driver will not yet use ANSI syntax for outer joins Date: Thu, 8 Aug 2002 17:23:01 +0200 Created on: DevCorelli Type: bugfix Problem Description (Technical terms) When using Oracle 9 and the resource ansi_outer_join enabled: Error ORA-01041 and ORA-04031 when ansi syntax is generated for outer joins. These errors do not occur when the Oracle proprietary syntax is generated. The mentioned Oracle errors indicate that there are memory management problems when the ANSI syntax is generated. This will cause problems at customer sites. Until the cause is known and the problem is solved (probably in a higher Oracle version), the oracle driver should not generate the ANSI syntax, but the old Oracle proprietary syntax for outer joins. Workaround set resource ansi_outer_join:0 Test Procedure PBC benchmark should run without errors MaintReger: # 15035 (BDUX8593): Wrong footer positioning when using Multi Column in Detail layout Date: Wed, 7 Aug 2002 16:11:15 +0200 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) A customized report is created with several header, detail and footer sections. In the detail section the multicolumn option is used. The last footer section should be printed only on the last page and at the bottom of the page. When the report is printed the footer is printed directly after the last detail line. Problem Description (Technical terms) When the report is compiled, where repgen is used, the setting of the variable lattr.multicol is set to 1 (horizontal), 2 (vertical) or 0 (= zero, no multi columns). The default setting is 0 (zero). In the section the lattr.multicol variable is set to 0 (zero) again. In the old situation this was done within a if statement, resulting in the fact that in some situations the variable was not set to zero. In the new situation the setting to zero is moved outside the if statement. Headers, details and footers are now printed correctly. Workaround Do not use multi columns. Test Procedure On the internal test system Cougar a report is created (tdinv342001960). When multi columns are used in the detail section and the report is compiled the footer.20 must be printed only on the last page and at the bottom of the page. On every page the header.30 should be printed. Motive source DF:123466 MaintReger: # 15005 (BDUX8591): use old javac to generate jar files Date: Tue, 6 Aug 2002 13:24:15 +0200 Created on: MaintReger Type: porting specific change Problem Description (Technical terms) Compilation error, because /usr/j2se/bin/javac is not on portingsystem. We put JDK1.3 header files in ClearCase and want to use javac 1.1 for generating jar files. Only JDK1.1 is needed on buildsystem. Test Procedure make bjvmi.jar MaintReger: # 14974 (BDUX8557): JDK headerfile path Date: Mon, 5 Aug 2002 10:49:15 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) When thirdparty vob is not mounted, wrong header files are included, because of fallback to local include dirs. Workaround Mount thirdparty vob. MaintReger: # 14934 (BDUX8519): merge of BDUX8466 Date: Tue, 30 Jul 2002 15:15:35 +0200 Project: BDUX8466 Problem Description (Customer terms) Error 254: Commit transaction has failed in the audit server. The following error is displayed in the message dialogue of BW: Error: ttaud.comt not according to MUST_BE_FILLED (value 0) Insert into ttaud110 compnr 000 failed. Can't create transaction notification ABORT Problem Description (Technical terms) The problem can be reproduced by doing a clear table w/o rollback followed by a commit on an audited table. See test procedure. Because the database drivers and audit servers did a low level internal commit after a clear_table_w/o_rollback, the status information of the transaction was already reset, while the audit information was not yet handled. This audit info was retrieved during the following commit, but because of the reset of the transaction status an empty GATI and Commit_time was inserted in ttaud110, violating the domain constraints of this table. Problem Description (Technical terms) (Required to fill this in) Workaround Disable the auditing on a table, but that might not be possible. Otherwise modify the session code: replace db.clear.table( , NO.ROLLBACK) by db.clear.table( , WITH.ROLLBACK) Workaround (Fill this in) Test Procedure Test the following 3GL code db.clear.table(ttaad050, NO.ROLLBACK) commit.transaction() or gcommand script #1 ttaad050 TCLTBL 1 #1 ttaad050 DCOMMIT 1 When AUDIT has been enabled for the table, error 254 occurs for the current porting set, while the problem is fixed in this project. Please verify the contents of the audit files and table ttaud110 as well. Test Procedure (Fill this in) Motive source DF:208440 Motive source DF:208440 MaintReger: # 14914 (BDUX8438): Scopus defect 126175: maximum size for referenced strings is not returned by xmlDllSignature Date: Fri, 26 Jul 2002 16:15:29 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) When generating a BOI with BaanDLL methods using BOI Modeler, and using a referenced string (inout parameter) for the BaanDLL, the maximum length of the outgoing string is the same as the length of the incoming string. Problem Description (Technical terms) The Bshell does not return the maximum string length when a ref string argument is specified in a DLL. Because this information is not supplied, the generated QCD/DCD files will also not contain a size specification. The end effect is that for ref string parameters, the size of the input string is used to allocate the ref string symbol within the Bshell Motive source Scopus:126175 MaintReger: # 14910 (BDUX8041): database server termination error Date: Fri, 26 Jul 2002 12:49:42 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) When running the session cprrp1210m000 by job, with multiple bshells and after restart of baan, the system comes up with detected database server termination. Problem Description (Technical terms) In some cases it could happend that the pending dd list has duplicate entries for the same dd. This caused a deep recursion in function detect_really_pending _dd(). The consequence is a stack trace overflow and a core file is created. This is a general problem and is not related to jobs. Workaround n/a Test Procedure Test is difficult. You need a system with this kind of error. Tests was done on three different systems. Result - no database termination anymore. Motive source DF#125112 MaintReger: # 14905 (BDNT8464): Fixed object is newer that source in Baan Debugger (when Baan is running on Windows NT) Date: Fri, 26 Jul 2002 12:37:44 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) Originally: solve a connection lost problem when debugging in the Baan Debugger when Baan is on Windows NT. Actually: fixed message in Baan debugger: object is newer than source. Problem Description (Technical terms) see above. The connection lost problem was already solved within project BDNT8265 (as side-effect). Test Procedure Compile a debug baan object and check that you don't get the message anymore. Motive source TCS:100-127864 (not true: BDNT8265 fixed this defect, this project fixes something else, see above) MaintReger: # 14894 (BDUX7888): Fix BCLM Client on IBM OS390 memory init bug. Date: Thu, 25 Jul 2002 17:18:46 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) BaanERP porting set will dump core when trying to run an OpenWorld (version 2.2 or higher) session. Problem Description (Technical terms) OpenWorld 2.2 uses the BCLM OpenWorld Client to request a license to run at the BCLM License Server. The OpenWorld session uses the bshell JVMI integration to be able to load the BCLM OpenWorld Client library. Inside this BCLM OpenWorld Client an IBM OS390 specific memory allocation initialisation is being called that may only be performed once per process. Because the same memory allocation initialisation has been performed inside the bshell also, OS390 will terminate the process. Test Procedure 1) Connect BaanERP menu browser to a bshell running on an IBM OS390 machine. 2) Start an OpenWorld (2.2 or higher) session (eg. tmboaserver). 3) If session starts, things work okay. MaintReger: # 14892 (BDUX8326): Fix memory allocation core dump on IBM OS390 when used with BCLM OpenWorld Client. Date: Thu, 25 Jul 2002 17:17:19 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Bshell dumps core when starting OpenWorld (version 2.2) session on an IBM OS390 machine. Problem Description (Technical terms) OpenWorld 2.2 uses the BCLM OpenWorld Client to request a license to run at the BCLM License Server. The OpenWorld session uses the bshell JVMI integration to be able to load the BCLM OpenWorld Client library. Inside this BCLM OpenWorld Client an IBM OS390 specific memory allocation initialisation is being called that may only be performed once per process. Because the same memory allocation initialisation has been performed inside the bshell also, OS390 will terminate the process. Test Procedure 1) Connect BaanERP menu browser to a bshell running on an IBM OS390 machine. 2) Start an OpenWorld (2.2 or higher) session (eg. tmboaserver). 3) If session starts, things work okay. MaintReger: # 14883 (BDNT8445): Fixed repgen error Date: Thu, 25 Jul 2002 10:44:12 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) Compiling reports on NT always results in a fatal error of repgen. Problem Description (Technical terms) Repgen was using unitialized data, causing strange results (i.e. corruption of the stack) when doing strcpy() (since the involved string was not \0 terminated). Although not reported, Unix is also involved with this problem. Test Procedure Compile a report on NT, e.g. in ttadv3530m000 Motive source TCS:106-125025 MaintReger: # 14872 (BDNT8461): Multibyte characters in pathname of .bwc file Date: Wed, 24 Jul 2002 16:25:09 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) 1. Take a bw configuration file. Copy and put it in a different folder. Now specify this one during the startup of the bss client. BSS client would be started now. 2. Now rename the folder name to some character in MB Character.(Something like C:\kums\‰@‰\‰|‰~ . Now specify this bw configuration file. You will get a message 'Invalid configuration file" and it is not possible to use that BW file. So we can confirm that the MB characters in the folder path of the bw configuraiton file is not supported. Problem Description (Technical terms) Multi-byte conversion was not done correctly in bwc.dll Workaround Don't store .bwc files in a path containing multibyte characters. Test Procedure Install BW in a location where multibyte characters are part of the pathname. Motive source TCS:601-125236 MaintReger: # 14847 (BDUX8020): Make use of new SUN Forte Developer 6 upd 2 compiler Date: Tue, 23 Jul 2002 13:29:48 +0200 Created on: MaintReger Type: porting specific change Problem Description (Technical terms) Added support for new SUn Forte 6 upd 2 compiler Test Procedure check whether code is compilable with this compiler MaintReger: # 14837 (BDUX8278): Segm violation caused by specific shm layout Date: Tue, 23 Jul 2002 09:18:22 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) On Unix platforms: core dump in the oracle driver with a stack trace like . . . . ( 4) 0x0011a064 core_dumped + 0x3c ( 5) 0xc01f2ee8 _sigreturn ( 6) 0xc01b8a00 memcpy + 0x1b8 ( 7) 0x0010b988 mal_memsave + 0x50 ( 8) 0x00053cac dbsCopyShmToDbsDD . . . . Problem Description (Technical terms) Note that this problem only occurs on Unix where multiple shared memory segments can be attached. On Windows platforms there is only one shared memory segment, so the problem does not occur. This crash is caused by reading more data than was allocated. Normally this would not be noticed as these extra bytes reside in memory that can be read by the driver. On a test system we had a situation where the read data was located at the very end of a shared memory segment, while the following segment was not attached at the next memory address. There was a gap in the memory space. This caused the driver to read the extra data from this gap, so from a memory location that was not attached to the process resulting in SIGSEGV. Workaround restart the BSE environment to clear the shared memory cache. The chance that the problem happens again is less than 0.01% per restart. Test Procedure Very hard to setup a test scenario as it depends on the actual layout of the virtual memory and the shared memory filling. Motive source TCS:106-126434 MaintReger: # 14833 (BDUX8282): Solve remark on no ubsp file available during DB install on AS400. Date: Mon, 22 Jul 2002 20:13:50 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) 1) The ubsp file was missing when db4_maint6.2 program was called from db4_adm6.2. 2) The installation was hanging at 72% of the import process. Problem Description (Technical terms) 1) The db4_maint6.2 program checks the u file at the beginning to collect user information. Probably because of timing, the file in some cases was missing from $BSE/lib/user. 2) The sort sequence table was not correct, the space was still converted to ASCII space, while we want it to be 0x40 (EBCDIC space). Workaround 1) Copy u file from an existing environment and restart the installation. 2) None. Test Procedure Run the Reger BaanERP installation on AS/400 system. Motive source TCS:100-127694 MaintReger: # 14818 (BDUX8407): LTS: TSSwc_width error about sign extended char Date: Fri, 19 Jul 2002 16:24:42 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) at some point simple characters above the ASCII range were sign extended, causing errors like "Can't convert TSS character 0xfffffffc" Test Procedure see TCS Motive source TCS:106-126109 MaintReger: # 14785 (BDUX8389): Bic lex and yacc files for AS400 (BDUX8329) bis Date: Thu, 18 Jul 2002 12:05:41 +0200 Created on: MaintReger Type: porting specific change Problem Description (Technical terms) File bic_tokens_aix.h was not renewed Test Procedure Nightly build on dilbert and frite MaintReger: # 14773 (BDNT8353): SHM on Windows: take case insensitivity of filenames in account Date: Wed, 17 Jul 2002 16:27:30 +0200 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) Only applicable for Windows platforms: In spite of the fact that tabledefinitions are loaded into the Baan Shared Memory, they are still read from disk. Because of this startup of a session is slow. Problem Description (Technical terms) There is a difference in lower/uppercase in the pathname of the stored tabledefinition, in comparison with the the pathname searched for. E.g.: The tabledefinition is stored in SHM as: C:\baan\dict\dttadv\dttadv999, and the client (bshell) is searching for: c:\baan\dict\dttadv\dttadv999. Because the first character "C" is stored as a capital in SHM, it cannot find the tabledefinition in SHM, and therefore loads it from disk. Workaround Take care the value of the BSE variable does not contain capital letters. (BSE variable is stored in registry: HKLM\Software\Baan\) Test Procedure 1. Set BSE value in registry to e.g. C:\baan (so capital C!) 2. Start the Baan Shared Memory service 3. Before logging in Baan specify the following in the Command field of BW Configuration: -- -dbgsrdduse -logfile shmtrace.log -keeplog 4. Log in Baan 5. Check out the file shmtrace.log, it should show the tabledefs are read from SHM (srdd). Motive source DF:125547 MaintReger: # 14770 (BDUX7553): BCLM: Fix bclm ID generation on Dec OSF. Date: Wed, 17 Jul 2002 15:47:22 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) Next error is logged on DEC OSF when startin BclmServer or BclmId. "Unable to retrieve cluster name" Problem Description (Technical terms) Inside the BclmServer and BclmId binaries a routine is being used to generate an unique hardware identifier. On DEC OSF clusters the cluster name is used in the calculation. The cluster detection process didn't work sufficient enough to detect a non-cluster environment. This caused the WARNING messages being logged on non-cluster DEC OSF environments. Test Procedure Run BclmId on non cluster DEC OSF system and see the difference. MaintReger: # 14756 (BDNT8364): BW does not always use the current date/time format Date: Wed, 17 Jul 2002 09:55:12 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) When computer date display settings are Dutch (today 12 july 2002 = 12/7/2002) then the date is not displayed correctly and there is a conflict between the dutch display format and the english internal format. Problem Description (Technical terms) System default locale is used in stead of user default locale to determine the default date/time format. Workaround Change default locale in the control pannel -> regional settings. Test Procedure See problem description above. Motive source TCS:106-126532 MaintReger: # 14754 (BDNT8344): Upgrade of thirdparty i386 DLLs Date: Wed, 17 Jul 2002 09:52:25 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Baan delivers the previously mentioned Microsoft DLLs as part of the BW application suite and the Baan delivered dlls are what BW is linking to on initialization. There are newer versions of these dlls that have been delivered by Microsoft and Baan needs to deliver the newest revisions of these dlls with the next released version of BW. Please validate that the most current versions of any Microsoft dll, being delivered as part of the BW app suite, are current and build this into the BW application build/delivery process. Problem Description (Technical terms) Refer to problem description. Upgrade dll version's in that way they are of the same version as delivered with Visual Studio 6.0 SP5. Refer to internet side http://support.microsoft.com/servicedesks/fileversion/dllinfo.asp?sd=MSDN for more info. Test Procedure n.a. Motive source DF:208886 MaintReger: # 14749 (BDNT8157): Enable ShowLightMargins for barcode type UPC-A Date: Tue, 16 Jul 2002 16:48:01 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) When printing barcode type 5, this is UPC-A, the most left and the most right digit is not printed. Problem Description (Technical terms) Option "ShowLightMargins" has to be enabled for this type of barcode to enable the mentioned digits. Workaround No workaround possible. Test Procedure Using bwprint.exe start the Barcode Test. When looking to barcode type 5, the number under the barcode image should start with digit 1 at the most left side, and end with digit 9 at the most right side. (so: 1 23487 17282 9) Motive source DF:11111 MaintReger: # 14734 (BDNT8306): Introduce no_sfw (No SetForegroundWindow) option Date: Tue, 16 Jul 2002 09:17:50 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) When running import or export exchange schemes with BW.exe at version 7.1c.02.143, the resultant ASCII information window remains on top of all other windows and is refreshed and forced on top with each change of the numbers on the display. The previous versions of BW.exe (such as 6.2a.02.xxx) only forced the window on top of its own BW.exe, not ALL windows. Problem Description (Technical terms) Like done already before on BaanIV (project BDNT5984) Workaround No workaround. Test Procedure Please open some application and minimize them. Log into the B50b Application Environment on the internal server called arnie. Go to session tuxch0233m000 "Export Data (Non-Regular). In this session fill in "focus2" for Exchange Scheme and "10" for Batch. Click on the button "Export Data". While the "Export Data" process is running; restore the minimized applications. You will see that as soon as an application is restored; the ASCII Window which shows the progress of the "Export Data" process (How many records have been processed) will "jump" on top of the application screen. Motive source DF:208357 MaintReger: # 14730 (BDNT8216): Color selection slidebar disappears below 100 and above 1 Date: Tue, 16 Jul 2002 09:06:05 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) The color selector in any Baan 5 environment, scrolling down from 255 to 99, removes the slide bar and moving up from 0, removes the slide bar. You can put a number in manually in the color field, for example, in session tgbrg9175s000 and then select the arrow and you can move a little futher on the bar. Problem Description (Technical terms) Problem caused geometry management (width of slidebar becomes 0, causing the slidebar to disappear). Workaround Put a number in manually in the color field. Test Procedure 1) Go to session tgbrg9175m000 2) Double click on a category to select. You will get tgbrg9175s000 3) Click on the arrow at the "color" field. You will get the color selection screen. 4) On the color selection screen, Move the indicator arrow down the bar from 255 and up from 0. You can also use session ttdsk3160m000 to start a ottdskcs. Motive source DF:207951 MaintReger: # 14723 (BDNT8340): Merged changes from project BDNT8015 to Reger and up Date: Mon, 15 Jul 2002 12:28:02 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) see project BDNT8015 Test Procedure run baanman.msc and check that baanman works well (no crashes, SHM configurable when SHM is not running). The changes for CreateProject to use less resources in a Windows Deskop are quite not testable. MaintReger: # 14709 (BDUX8329): bic6.X generate core dump during compilation Date: Fri, 12 Jul 2002 11:56:15 +0200 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) bic6.2 produce core dump during compilation. function main() { test(ltoe(1)) } function test( domain tcyesno yesno) { } bic6.2 script -u -o myobj -f tmp1 Problem Description (Technical terms) The problem appears when -u option is used only. The core file will be created in function check_fun_arg() and is related to the function ltoe() used in the script. The statement where the core dump will be created is if( ptr && (ptr->flags & TABLE) ) { In case of function ltoe() is used in a script the pointer ptr has a invalid value (0x1). Workaround n/a Test Procedure bic6.2 script -u -o myobj -f tmp1 Motive source DF:126659 MaintReger: # 14684 (BDNT8305): Merge BDNT8226 Date: Wed, 10 Jul 2002 13:24:04 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) BDNT8226 was not merged to MaintReger and DevEscher. Workaround n/a Test Procedure n/a MaintReger: # 14682 (BDNT7070): cleanup bw.bld Date: Wed, 10 Jul 2002 13:22:45 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) Cleanup of old installer stuff, which is not used anymore. MaintReger: # 14677 (BDUX8297): remove -xdepend flag Date: Tue, 9 Jul 2002 16:47:49 +0200 Created on: MaintBaanIVc Type: porting specific change Problem Description (Technical terms) FD6upd2 compiler dumps a core when sort.c is compiled. Workaround Use old WS5 compiler MaintReger: # 14668 (BDUX8209): bdbpost with option append is very slow Date: Tue, 9 Jul 2002 13:34:35 +0200 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) bdbpost with option append is very slow. Problem Description (Technical terms) The problem appears if ora_max_array_insert is set to value greater 1 and a lot of duplicate values occured during insert records. The insert buffer is not reseted. Workaround use ora_max_array_insert:1 in db_resource file. Test Procedure bdbpost6.1 -A -m -n < dump Motive source DF:125454 MaintReger: # 14665 (BDUX8289): Increment BW number due to rebuild of Reger SP1 BW Date: Tue, 9 Jul 2002 11:21:40 +0200 Created on: MaintReger Type: porting specific change Problem Description (Technical terms) Version number update Test Procedure check help->about in becs/BW MaintReger: # 14663 (BDUX7918): Merge project bugf_BDUX7906 Date: Tue, 9 Jul 2002 11:19:34 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) Merge from project bugf_BDUX7906 MaintReger: # 14661 (BDUX8279): Fixed compilation error due to new compiler for SUN (WS6upd2) Date: Tue, 9 Jul 2002 11:15:03 +0200 Created on: MaintReger Type: porting specific change Problem Description (Technical terms) Make compilation flags 'indentical' to corelli. Note: compilation of sort.o results still in a core dump if -xdepend is used (problem has been reported to SUN) Test Procedure build MaintReger: # 14654 (BDUX8283): Solved compilation problems in BDNT8265 Date: Tue, 9 Jul 2002 10:05:17 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) Fix BDNT8265 was not portable to non-Windows platforms. Test Procedure Compile! MaintReger: # 14650 (BDNT8265): Added support for table name aliasing in the from part of a BaanSQL call through OpenWorld Date: Mon, 8 Jul 2002 17:36:30 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) When using an OpenWorld query with table name aliases in the from clause, the query could not be executed. The same query however could be executed without any problems through the standard qptool command. An example of a query that failed is: select a.* from ttadv100 a Problem Description (Technical terms) The XML Query layer gets incorrect information from the QP layer. It gets the following info in the CRT_EXTVAR callback: - all column names, with the actual table name prepended (e.g. ttadv100.cpac) - the alias for the table name. Test Procedure Run the following query through OpenWorld: select a.* from ttadv100 It should succeed without problems. Motive source TCS:106-126351 MaintReger: # 14647 (BDUX8241): NQP: distributed subquery with 2 outer references loops Date: Mon, 8 Jul 2002 17:31:14 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) If a query is handled by two database drivers, the bshell may loop infinitely or the database driver may core dump. The bshell loops if a subquery contains two or more references to different columns of a table that is defined outside the subquery. Example: select empno from dbtst120 where exists ( select * from dbtst190 where empno = dbtst120.empno + dbtst120.sex ) The database driver dumps a core if a subquery contains a single reference to a column of a table that is defined outside the subquery. Example: select empno from dbtst120 where exists ( select * from dbtst190 where empno = dbtst120.empno ) Note that the problem only occurs if dbtst120 and dbtst190 are handled by different (instances of) database drivers. The transformation of outer column references into parameters entered an infinite loop, if more than one outer reference was present in the subquery. The network packet created for an outer reference was not correct, which resulted in a core dump of the database driver. Test Procedure Let dbtst120 and dbtst190 be handled by different database drivers. For example, tabledef6.2 may have the following entries: dbtst120:*:db2v5(DB2INSTANCE=db2inst1,X):N *:*:db2v5(DB2INSTANCE=db2inst1):N Execute the example queries in the "Problem Description (Customer terms)" section. No infinite loop should be entered. No core dump should be generated. The result should be correct. MaintReger: # 14637 (BDUX8237): Fixed compilation error on AS/400 and OS/390 caused by BDUX8145 Date: Fri, 5 Jul 2002 17:14:08 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) Fixed compilation errors on AS/400 Test Procedure compile MaintReger: # 14633 (BDNT8097): Add bshell functions bms.init and ams.control Date: Fri, 5 Jul 2002 17:09:46 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Problems all applicable when logging in Baan using Worktop. bms.init: Need to be able to initialize bms from 4GL program (refer to defect 207469) ams.control: In case authorisations are set via DEM, and via AMS, we need to be able to switch off AMS in case we are working in DEM. Now AMS authorisations are not switched off, when logging in Baan using Worktop, and having the DEM Browser specified as Startup Program for BW. The DMB (DEM Browser) is started when logging in using Worktop: Because Worktop has its own integrated Baan Process Browser, the DMB should not be started. This problem is introduced by project VHUX2585. Problem Description (Technical terms) bms.init: Bshell function bms.init needed ams.control: Bshell function ams.control needed The DMB (DEM Browser) is started when logging in using Worktop: There is no check done whether you are logged on via Worktop or another user interface. In case of Worktop, tgbrg4260m000 should not be started. Workaround bms.init: no workaround ams.control: no workaround The DMB (DEM Browser) is started when logging in using Worktop: Do not specify the DEM Browser as Startup Program for BW, in the User Data. Test Procedure Now when logging in Baan using Worktop, the bshell should not start the DEM Browser. (User Data, Startup Options: BW Startup Program = DEM Browser, Worktop Tree enabled, Options: Baan Process Browser) Test different kind of scenarios where (conflicting) authorisations are set in DEM, and in AMS. When logging in Baan using Worktop, the AMS authorisations should be applied ONLY in case you the Baan Menu Browser exists in Worktop. In case you have the Baan Process Browser in Worktop, the DEM authorisations should only apply. Motive source DF:124131 DF:207469 MaintReger: # 14606 (BDUX8251): Informix:Fix for error 1210 in import:TCS-100/127588 Date: Thu, 4 Jul 2002 09:21:48 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) Not Applicable Problem Description (Technical terms) The Baan min date is an out of range value for Informix and hence error -1210 is returned, if a date column is assigned the min date value. Test Procedure Create a table with date column (ttadv998 is used here as an example). Run the following gcommand script #1 ttadv998 RINS 0 -i -p = ttadv998.date = 0 = ttadv998.time = 0 = ttadv998.name = tttadv111 #1 ttadv998 DCOM 0 With a driver that does not contain this fix error -1210 will be returned on DEC. With this fix, the insertion would go through successfully. Motive source TCS:100-127588 MaintReger: # 14584 (BDUX8221): Fixed compilation errors on AS/400 and OS/390 introduced by project BDUX8145 Date: Tue, 2 Jul 2002 17:30:29 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) Fixed OS/390 and AS/400 compilation errors. Test Procedure Build asm_glob object (in lib/nw_1) and al_sit objet (in lib/al_1). No compilation errors should occur. MaintReger: # 14579 (BDNT8219): Check for local admin in BW setup not correct Date: Tue, 2 Jul 2002 14:56:19 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) Cannot install Baan Windows on a NT system other than a English version of Windows. The following error message is given: "You must be an administrator to install this software". Problem Description (Technical terms) The check for local administrator is language dependant. Test Procedure Install Baan Windows on a non-english version of NT (for instance Dutch). MaintReger: # 14569 (BDUX8205): SQL: parser accepts host variable 1.assembly.line Date: Tue, 2 Jul 2002 09:41:36 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) My script compiles succesfully, but when I run it, I get SqlState=HYS05, 'Parameter '1.assembly.line' not bound before exec' The parameter '1.assembly.line' starts with a 'one', so it cannot be a valid program variable name. Problem Description (Technical terms) The SQL parser accepts parameter names that start with a digit. Test Procedure Execute the following command: bsql6.2 -a p -q "select * from dbtst120 where empno = :1.empno" Verify that it returns: SqlState=42I00, 'Unexpected token '.'', NativeError=302, Line=1 Motive source TCS:106-122504 MaintReger: # 14567 (BDUX8212): compilation error AIX Date: Tue, 2 Jul 2002 09:38:33 +0200 Created on: MaintBaanIVc Type: porting specific change Problem Description (Technical terms) Undefined PJNICALL Test Procedure Compile MaintReger: # 14562 (BDUX8208): MaintReger: compile error on SUN Date: Tue, 2 Jul 2002 09:36:09 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) /vobs/tt/lib/al_1/alloc.c does not compile when compiled with -DOVERLEAP. Test Procedure Compile al_alloc.c with -DOVERLEAP. MaintReger: # 14560 (BDUX8193): JDK1.3 support Date: Mon, 1 Jul 2002 15:16:15 +0200 Created on: MaintBaanIVc Type: porting specific change Problem Description (Customer terms) Can not use JRE1.3 on Solaris to load libjava.so Problem Description (Technical terms) Bshell uses old way of JNI_CreateJavaVM. In JDK 1.1, the second argument to JNI_CreateJavaVM is always a pointer to JNIEnv *. The third argument is a pointer to a JDK 1.1 specific structure (JDK1_1InitArgs). The JDK1_1InitArgs structure is clearly not designed to be portable on all VMs. In the Java 2 SDK, we introduce a standard VM initialization structure. Backward compatibility is preserved. If the VM initialization argument points to a JDK1_1InitArgs structure, JNI_CreateJavaVM still returns the 1.1 version of JNI interface pointer. The VM returns the 1.2 version of JNI interface pointer if the third argument points to a JavaVMInitArgs structure. Unlike JDK1_1InitArgs, which contains a fixed set of options, JavaVMInitArgs uses option strings to encode arbitrary VM start up options. Though it should be backwards compatible, the HotSpot VM of JRE1.3 on Solaris does not support the JDK1_1InitArgs structure. Workaround Use JRE 1.1.8 or 1.2.x Test Procedure Run tmboaserver on a platform with JRE1.3. Create a file $HOME/jvm_options or $BSE/java/jvm_options. When $HOME/jvm_options is found $BSE/java/jvm_options will not be read. In that way a certain user can use its own options. $ java -help or java -X shows which options are availlable. For example: put in jvm_options: -Xmx=128000000 -hotspot MaintReger: # 14552 (BDUX8171): Some minor fixes in ASM, e.g. usage and some (debug) logging. Date: Mon, 1 Jul 2002 14:47:30 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) Fixed usage text and logging text (just only textual changes) Test Procedure Windows NT only - run asm_srv -?, and chech that -start|install|remove|stop is within {} instead of [], since the argument is mandatory (on Windows NT) - whith debugging (-d flag) enabled, check that environment are shown without trailing spaces MaintReger: # 14550 (BDUX8206): platfrom specific java_integr.c Date: Mon, 1 Jul 2002 14:24:57 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) Compilation error bshelljava on HPUX No JVMI support on SGI MaintReger: # 14548 (HYNT8127): MSQL:cannot open version info:TCS 125883 Date: Mon, 1 Jul 2002 13:09:59 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) The system information tab in the Baan About box when clicked, generated fatal error for MSQL Driver. Down the line, there is Null pointer assignment which causes the error. This problem occurs in NT Build. For oracle, it was solved in the database dependent layer. Test Procedure Start a reger environment using Worktop. When the baan icon appears on system tray, right click on it and select Help About. In the about screen click, System Info... You should be able to see the system information. Motive source TCS:100-125883 MaintReger: # 14547 (BDUX8183): TSS in query string literals Date: Mon, 1 Jul 2002 12:26:50 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) ittstpqbf does not work if kanji "Harai" is included in condition. Problem Description (Technical terms) The query parser does not recognize TSS characters in string literals. As a result of this the query parser will return an error if a string literal contains a TSS character that contains the same quote as is used for the string literal. (A string literal can have single or double quotes). The kanji Harai character is such a TSS character. This character is represented in TSS as: 0x9b, 0x21, 0x4a, 0x27. 0x27 is a single quote. Workaround Escape the quote in the TSS character. For example: In case of the Harai character, insert an extra 0x27 character in the TSS sequence and the query will work fine: 0x9b, 0x21, 0x4a, 0x27, 0x27. Motive source DF:119923 MaintReger: # 14544 (BDUX8197): DB: merge of BDUX7878: memory leak with auditing Date: Mon, 1 Jul 2002 11:59:32 +0200 Created on: DevEscher Type: bugfix Problem Description (Customer terms) "BW exits when we try to create very large number of records in Baan using a BOI." The insertion of a Purchase Order leads to starting a new session in the bshell (function server), the insertion of records, and the ending of the session. For each Purchase Order this procedure is followed. Each time a session is ended, an amount of memory is not freed. When a large amount of Purchase Orders is inserted, the bshell will abort because of memory exhaustion. This only happens when auditing is enabled on the table on which the inserts are done. Workaround If possible, disable auditing. Test Procedure Use the following test scripts (Baan3gl): Source t.c: function main() { long i, pid for i=1 to 1000 pid = activate("./s.o") wait( pid, WTHANG ) endfor } Source s.c (if needed, modify the source to use another table): table tdbtst120 function main() { error.bypass = 1 dbtst120.empno = 17 dbtst120.workdept = "A00" db.insert( tdbtst120, db.retry ) db.first( tdbtst120 ) abort.transaction() } Compile the scripts: bic6.1 ./t.c -o ./t.o bic6.1 ./s.c -o ./s.o Run the object t.o: bshell6.1 -server ./t.o Now observe the size of the bshell process (e.g. using ps -l). After an initial growth, it should very quickly stabilize. Motive source DF:124015 MaintReger: # 14537 (BDNT8151): BW access violation in rowcolumn widget Date: Mon, 1 Jul 2002 11:29:54 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) An error takes place at various times while BW is running WinRunner scripts. A message box is received with the title: tcibd0101s000:General Item Data (675) dcaca314:bw.exe - Application Error. The contents of the message box reads "The instruction at "0x5f4012a1" referenced memory at "0x00000004". The memory could not be "read". Click on OK to terminate the program." Additional Information: The error regarding "referenced memory ... The memory could not be "read"" creates the following message in the NT event log: Event Type: Error Event Source: Baan Event Category: (7) Event ID: 1000 Date: 6/17/2002 Time: 6:42:31 AM User: NW\trnxrun1 Computer: E120677 Description: The description for Event ID ( 1000 ) in Source ( Baan ) cannot be found. The local computer may not have the necessary registry information or message DLL files to display messages from a remote computer. The following information is part of the event: Env Bse (\\dcac-nt-10\dcac_develop\baan\bw) Prog BW file \general\lib\al_1\al_wutil.c # 534 Keyword Stack trace Username trnxrun1 type S language 2 Process 0x54c ProcessSet trnxrun1@E120677:1356 Lasterror 0 Errno 0 bdberrno 0 Message Exception C0000005 (Access violation) TsRowColWidgetRec::changeLayoutPE+1C8 (1:000a9358) TsRowColWidgetRec::changeLayout+45 (1:000a8bf5) TsRowColWidgetRec::changeManaged+5E (1:000a838e) TsCompositeRec::doChangeManaged+4C (1:000b9c6c) TsCompositeRec::doChangeManaged+33 (1:000b9c53) TsCompositeRec::doChangeManaged+33 (1:000b9c53) TsCompositeRec::doChangeManaged+33 (1:000b9c53) TsWidgetRec::topRealize+16 (1:00077736) TsToolkit::realizeWidget+B (1:0006deab) TsRealizeWidget+12 (1:00068112) DmRealizeMWindow+15 (1:000633b5) DmCheckObjectResourceTable+2584 (1:0005d914) DmFreeArgList+46F (1:0006170f) DmChangeCompoundWidgetProc+13 (1:00060e93) DmChangeWinObject+125 (1:0005fc85) DmRequestTimerHandler+AA (1:00056baa) DmRequestInputHandler+5 (1:00056ac5) TsInputIdRec::streamCallbackProc+30 (1:000b9080) SocketData::processStreamData+1B (1:00022d3b) SocketData::processIncomingData+C9 (1:00022e29) SocketData::windowProc+1BD (1:0002306d) Problem Description (Technical terms) In the rowcolumn widget there is an error in the algorithm causing sometimes a memory read beyond an allocated array. Test Procedure This failure only occured after extensive testing with WinRunner scripts. It probably occurs when a 4GL Find dialog is started. Motive source DF:208551 MaintReger: # 14523 (BDNT7800): Enhanced logging in ASM Date: Fri, 28 Jun 2002 15:25:47 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Enhanced logging of asm_cli utilty, and in some cases of ASM server. Problem Description (Technical terms) Provide better diagnostics in case of troubles with ASM with asm_cli utility. MaintReger: # 14511 (BDNT8176): Merge from BDNT8090 to Reger and Gemini Date: Fri, 28 Jun 2002 10:10:09 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) >From the BW login dialog you can press button Configure to come into the BW Configuration. In the BW Configuration sheet you have the following buttons: OK Cancel Apply When modifying the configuration, eg. changing the password, the Apply button gets enabled. When pressing the Apply button, only the modifications for this config sheet object are saved, so NOT saved to disk (bwc file). When pressing now the OK button, the program (bwc.dll) tries to save the modifications to the runtime bwc file. In case this file is read-only, this happens in central deployment environment, the following happens: You get the following error from BW: "Could not save configuration file ".............bwc" The solution is twofold: - Don't shown the Apply button anymore in the config dialog - Don't try to save a configuration object anymore when in central deployment mode. Problem Description (Technical terms) See problem description above. Workaround Just ignore. Test Procedure See problem description. Motive source DF:207427 MaintReger: # 14505 (BDUX8166): Partly merged project BDUX8144 to Reger and higher Date: Thu, 27 Jun 2002 16:34:45 +0200 Created on: MaintReger Type: porting specific change Problem Description (Technical terms) Made Reger code compilable for SUN Forte Developer 6 upd 2 compiler Workaround nona Test Procedure Compiler MaintReger/Gemini on Solaris with WS6upd2 compiler MaintReger: # 14493 (BDUX8145): Fixed error in get_version() Date: Thu, 27 Jun 2002 09:47:24 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Customer terms) System information causes core-dump in bshell on Windows NT. Problem Description (Technical terms) Do not return NULL pointers with get_version() (which is used in strcpy() etc for system info). This problem only occurs on NT. Test Procedure Run session ttsit0100m000 Motive source TCS:100-125883 MaintReger: # 14489 (BDUX8162): Merge bugf_BDUX8154 Date: Wed, 26 Jun 2002 17:01:01 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) Merge bugf_BDUX8154 MaintReger: # 14479 (BDUX8081): JVM startup arguments Date: Tue, 25 Jun 2002 17:05:18 +0200 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) Can not use JRE1.3 on Solaris to load libjava.so Problem Description (Technical terms) Bshell uses old way of JNI_CreateJavaVM. In JDK 1.1, the second argument to JNI_CreateJavaVM is always a pointer to JNIEn v *. The third argument is a pointer to a JDK 1.1 specific structure (JDK1_1Init Args). The JDK1_1InitArgs structure is clearly not designed to be portable on al l VMs. In the Java 2 SDK, we introduce a standard VM initialization structure. Backward compatibility is preserved. If the VM initialization argument points to a JDK1_ 1InitArgs structure, JNI_CreateJavaVM still returns the 1.1 version of JNI inter face pointer. The VM returns the 1.2 version of JNI interface pointer if the thi rd argument points to a JavaVMInitArgs structure. Unlike JDK1_1InitArgs, which c ontains a fixed set of options, JavaVMInitArgs uses option strings to encode arb itrary VM start up options. Though it should be backwards compatible, the HotSpot VM of JRE1.3 on Solaris do es not support the JDK1_1InitArgs structure. Workaround Use JRE 1.1.8 or 1.2.x Test Procedure Run tmboaserver on a platform with JRE1.3. Create a file $HOME/jvm_options or $BSE/java/jvm_options. When $HOME/jvm_options is found $BSE/java/jvm_options will not be read. In that way a certain user can use its own options. $ java -help or java -X shows which options are availlable. For example: put in jvm_options: -Xmx=128000000 -hotspot MaintReger: # 14471 (BDUX8109): RSTFRMSTMF has a service program bind to it with fixed libaray name. Date: Tue, 25 Jun 2002 13:51:45 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) RSTFRMSTMF command, which part of the portingset does not work on V5R1. Problem Description (Technical terms) During the build of the command processing program of RSTFRMSTMF service program WRKO4OBJS is bind with fixed library name, which is the build library. Since this library is not created on the install system the command did not work. Workaround Create the build library on the system and copy service program WRKO4OBJS to this library. The build library can be determined by 'DSPPGM PGM(/WORK04OBJS) DETAIL(*SRVPGM)' Motive source DF:208521 MaintReger: # 14460 (BDUX8067): uni.import function fix Date: Mon, 24 Jun 2002 15:29:17 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) Two different bugs: 1. bshell crashed when uni.import() is gets data from an odd length file for converting. 2. At an invalid character it gived back -1 expect -3 as a return value. Problem Description (Technical terms) 1. Misaligned data casted to SHORT type and tries to read it -> BUS error on HPUX systems. 2. Missed a check after searching from the table. Test Procedure See the TCS renmark for example sources. Motive source TCS:106-125952 MaintReger: # 14438 (BDUX7944): Oracle installer: XML file contains plain default password Date: Fri, 21 Jun 2002 11:28:32 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) The XML file generated during the Baan/Oracle installation contains a password in plain text. This happens for the publically known default value of the Oracle DBA password in the inactive option sets in the XML file. All actual passwords are encrypted, so the problem is more cosmetic than a real security risk. Test Procedure During an installation check the generated XML file. All passwords in the XML file should be encrypted. Motive source TCS:100-126426 MaintReger: # 14433 (BDUX7288): Oracle installer should default to UNDO tablespace for Oracle9 Date: Thu, 20 Jun 2002 17:02:18 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) 1) When an installation is done on Oracle9 the datafiles belonging to the TEMPORARY tablespace are not shown. When installing on Oracle8 there is no problem. 2) When an installation is done on Oracle9 the defaults are set to using RBS (Rollback segments) instead of UNDO tablespaces. When the default is choosen, this leads to errors later on during the installation when the RBS tablespace is extended by the installer. 3) The minimum value for the array interface is 2. This should be 1 to offer the possibility to enable one parameter while disabling the other. Problem Description (Technical terms) 1) The Oracle installer shows all datafiles that are attached to a tablespace. This information is retrieved from the DBA_DATA_FILES view. In Oracle9 the files containing temporary data (like all files of the TEMP tablespace) are not in this view anymore. For temporary data the view DBA_TEMP_FILES should be used. When connected to Oracle8 the installer behavior should not be changed. 2) When connected to Oracle9 the installer should default to UNDO tablespaces instead of RBS (rollback segments). RBS is legacy. When connected to Oracle8 the installer behavior should not be changed. Test Procedure Do an install on Oracle 9 and one on Oracle 8 for verifying the new behavior on Oracle9 and the existing one on Oracle8 Motive source TCS 100-127359 MaintReger: # 14430 (BDUX8093): Correct the badfix for bdbpost skip table Date: Thu, 20 Jun 2002 15:35:39 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Customer terms) Bdbpost gives the error message: "Can not read table definition (error 0)" Problem Description (Technical terms) The strings "#$tppmmmxxxccc" is stored in an array of chars that is too short. Bdbpost stops processing if loading of the data of one table fails. Workaround No workaround. Test Procedure Create with bdbpre6.2 an export file. For example: $ bdbpre6.2 -N dbhvl011 -C 700 > dbhvl011700.dmp Recreate the table with bdbpost from the export file on an AIX system. For example: $ bdbpost6.2 -k < dbhvl011700.dmp The output of the old and wrong bdbpost likes: Default separator ('\0') taken Table M Idx RowCount Messages 00700 Can not read table definition (error 0) The output of the new and correct bdbpost likes: Default separator ('\0') taken Table M Idx RowCount Messages dbhvl011700 1 400 Create with bdbpre6.2 an export file of three tables. The first table should contain at least 1Kbyte of data. For example: $ bdbpre6.2 -k -N dbhvl011 dbhvl012 dbhvl013 -C 700 > table.dmp Create in the "storage" file of the used database a seperate line for the data of the first table. For example: (Oracle): dbhvl011:700:T:TABLESPACE datspace PCTFREE 80 PCTUSED 20 INITRANS 3 \ STORAGE(INITIAL 1K MAXEXTENTS 1) The aim of this line is that loading the data in this table will fail very soon. Recreate the tables with bdbpost from the export file. For example: $ bdbpost6.2 -nf < table.dmp The output of the old and wrong bdbpost likes: Default separator ('\0') taken Table M Idx RowCount Messages dbhvl011700 1 100 Can not commit transaction (error 520) Bdbpost6.2 stops after the failure in the first table. The output of the new and correct bdbpost likes: Default separator ('\0') taken Table M Idx RowCount Messages dbhvl011700 1 100 Can not commit transaction (error 520) dbhvl012700 1 2 dbhvl013700 1 2 Bdbpost6.2 continues after the failure in the first table. MaintReger: # 14420 (BDUX8089): Oracle installer sets default tablespace for created users Date: Wed, 19 Jun 2002 17:52:31 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) Oracle users generated by the Baan/Oracle installer have 'SYSTEM' as default tablespace. Problem Description (Technical terms) Oracle is normally using the "SYSTEM" tablespace as default tablespace. However users are not supposed to (but nothing prevents this!) create tables in SYSTEM, to avoid problems (like full tablespace) that might cause the whole instance to run into problems. Also for maintenance reasons it is better to split oracle-internal stuff from application data. For the "Baan" user (owning all tables) the default tablespace is set to the tablespace that contains the BaanERP tables, although this setting will never be used: All create table/index commands generated by the Baan/Oracle driver have a TABLESPACE clause (derived from $BSE/lib/ora/ora_storage_param), so the default is not needed. For all other Oracle users created by BaanERP, the default tablespace is completely irrelevant: - All users (except the "Baan" user) do not have any permission to create any table in Oracle, so unless a DBA has manually changed the permissions in Oracle, no tables can be created by these users, and thus the default tablespace is never used. - All tables that are created by the Oracle driver are created by the group-user, never by users like "bsp". So in this case "Baan" creates and owns all tables and indexes. This user has a default tablespace set. To avoid questions because people only look at the default tablespace setting and start asking questions, it was decided to change the installer. Test Procedure Do a BaanERP5.2 install, and verify the created Oracle users. Motive source TCS:100-127546 MaintReger: # 14415 (BDNT8095): Completion message after upgrading BW Date: Wed, 19 Jun 2002 15:27:46 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) When updating a Baan Windows setup, the setup program exits without giving a message to the user. Problem Description (Technical terms) No completion dialog was presented after a sucessfull update of Baan Windows. Test Procedure Update a previous version of Baan Windows. Motive source TCS:100-127555 MaintReger: # 14384 (BDUX8006): OVERLEAP enable in bic Date: Mon, 17 Jun 2002 13:20:21 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) Bic doesn't made any logs in OVERLEAP mode. Workaround Put a calling for toggle_mal_log() into the bic main. Test Procedure Compile the bic with OVERLEAP defined, and set the enviroment variable also, then execute the bic. See the log file as a result. Motive source TCS:106-111860 MaintReger: # 14381 (BDUX8063): C-int merge fix for as400/os390 Date: Mon, 17 Jun 2002 09:58:35 +0200 Created on: MaintBaanIVc Type: porting specific change Motive source DF:207273 MaintReger: # 14376 (BDUX8049): Reduce BCLM logging. Date: Fri, 14 Jun 2002 16:48:11 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) Too much BCLM error messages are logged into the bshell log. Problem Description (Technical terms) See customer terms. Workaround Install BCLM Client for BaanERP on system will reduce the number of errors logged. Test Procedure Run bshell with and without BCLM Client for BaanERP installed on system. Motive source TCS:100-124913 TCS:100-124949 TCS:100-127466 TCS:200-73037 MaintReger: # 14368 (BDUX7985): Create fixes for DB2/400 regression tests. Date: Fri, 14 Jun 2002 14:15:17 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) The Reger DB2/400 installation and regression test does not run properly. During the installation the CRTBAANCOL command did not work unless the Baan library was added to the library list; the collection name could not contain underscore. The regression test returned wrong result sets. Problem Description (Technical terms) Installation: The in the CRTBAANCOL the COL parameter had wrong type, it didn't allow underscore. In the RESTOBJS program some of the commands were not changed to look for the programs in the Baan library. Regression test: The 'having count(*)' clause caused pointer not set error, because the input variable couln't be binded, it was not defined at that point - fix was copied from DB6000 code. If the input host variable was NULL, the data was not treated as NULL, because the db_type should be odd as well to indicate the NULL value (setting the NULL indicator ro -1 was not enough). The DB engine expands strings by spaces if one of them is longer if there is a character comparison in the query - select * from dbtst120 where firstnme > 'CHRISTINE'; domain lenngth 20 chars. The engine expanded it by 0x40 (EBCDIC space) but the host variable was translated to ASCII, so the spaces where 0x20. It will be solved in a later release, therefore a workaround in the code was needed. The sort sequence table was changed to transfer EBCDIC space (0x40) to EBCDIC space, while every other character is transfered to ASCII. Test Procedure Installation: run an installation in Reger. Regression test: run the regression tests. Motive source TCS:100-126682 MaintReger: # 14365 (BDUX8050): compilation error OS390 Date: Thu, 13 Jun 2002 16:56:01 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) Compilation error Test Procedure compile MaintReger: # 14348 (BDUX8042): non existing include files on OS/390 Date: Thu, 13 Jun 2002 11:27:38 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) On OS390 we get a message during compilation: WARNING CBC3296 /port.7.1c.04/vobs/tt/lib/hal/user.c:17 #include file not found. WARNING CBC3296 /port.7.1c.04/vobs/tt/lib/tax/taxpvtx/vtx_he/port.h:203 #include file not found. WARNING CBC3296 /port.7.1c.04/vobs/tt/mir/mir/mir_prtf.c:6 #include file not found. Workaround Ignore these WARNINGS, though it should be an ERROR! Test Procedure compile on OS390 MaintReger: # 14337 (BDUX8029): Fixed build problem on OS/390 Date: Wed, 12 Jun 2002 12:13:11 +0200 Created on: MaintReger Type: porting specific change Problem Description (Technical terms) Fixed build problem on OS/390 MaintReger: # 14329 (BDNT8008): Time control in grid returns wrong date part Date: Wed, 12 Jun 2002 08:47:03 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) Start session tcmcs0108m000 Now insert a record with a future date Now change the value in Time segment and not date segment, change this using the keyboard. Then observe that it is saving with a wrong date. Problem Description (Technical terms) In a time-only control, the date part of time was reset to a constant value when the time part was changed using the keyboard. Test Procedure See problem description above. Motive source TCS:123-12345 MaintReger: # 14327 (BDUX8026): C-Interface performance fixes Date: Wed, 12 Jun 2002 08:40:06 +0200 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) When using the Baan c-interface the created api can at times use up all (or most)of the CPU. Problem Description (Technical terms) When the api routine is waiting on a response from the Baan side it polls very rapidly for a reply. This polling causes very high CPU usage to occur. Test Procedure Create a test api which calls a Baan dll function. The baan dll function should have some long suspend()'s coded in them. This simulates a long response from the Baan side. Run new api i(built agains new libci.a lib) and observe that the new functionality reduces cpu usage significantly. Motive source DF:207273 MaintReger: # 14317 (BDUX7912): C-Interface Perf. Enhancement Date: Mon, 10 Jun 2002 11:54:05 +0200 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) When using the Baan c-interface the created api can at times use up all (or most)of the CPU. Problem Description (Technical terms) When the api routine is waiting on a response from the Baan side it polls very rapidly for a reply. This polling causes very high CPU usage to occur. Test Procedure Create a test api which calls a Baan dll function. The baan dll function should have some long suspend()'s coded in them. This simulates a long response from the Baan side. Run new api i(built agains new libci.a lib) and observe that the new functionality reduces cpu usage significantly. MaintReger: # 14308 (BDNT8003): Grid readonly cells can get focus Date: Fri, 7 Jun 2002 16:49:42 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) The new Date/Time control in the grid gets focus when the keyboard is used, while the control is readonly. See session: bptmm1130m000 When the first line cannot be edited and the seccond line can be edited. When focus is on column 0 of the grid, press Tab. You'll see that the grid scrolls, however, you can't see that focus is set on the Date part of the Transaction Date. But in the 4GL engine you can see it. Problem Description (Technical terms) A readonly date/time control in the Grid was still part of the Grid focus chain. Workaround Use mouse navigation. Test Procedure See problem description. Motive source TCS:106-125738 MaintReger: # 14300 (BDNT7993): lost connection Date: Fri, 7 Jun 2002 14:51:40 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) When installing reger on W2k / Oracle 9, Error during dba_processing occured. Connection to server lost: fd_read 508: num_bytes -1 errno 0 Problem Description (Technical terms) Investigation found out that closing sh_server by bshell give problems. sh_server gets QUIT command, sends a SH_STARTOK to bshell and exits. bshell is not able to read the SH_STARTOK in time and reads on a non-existing socket. Test Procedure Run BSI. Motive source TCS:100-127478 MaintReger: # 14298 (BDNT7994): Added OG as product for sending reports after a build Date: Fri, 7 Jun 2002 14:49:59 +0200 Created on: MaintReger Type: porting specific change Problem Description (Technical terms) Open Galaxy was not a known product for porting scripts for builds on NT Workaround N/A Test Procedure Run make_pbc_cs.bat and mak_report.bat and check that OG is selected as product. Prerequisites: - config spec has rule: element bproduct.h .../og/LATEST - general and ConfigTools vob are also available MaintReger: # 14294 (BDUX7999): seq.skip fix Date: Fri, 7 Jun 2002 14:47:33 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) seq.skip does not always work properly Problem Description (Technical terms) seq.skip's buffer handling method produce some random errors. Workaround Remapping seq.skip(n,fp) to seq.seek(n,1,fp) Motive source TCS:100-125536 MaintReger: # 14292 (BDUX7991): option 2 for open.message does not a correct check. Date: Fri, 7 Jun 2002 11:23:50 +0200 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) According to the Baan help, open.message(...) should return -1 and not create a new mailbox when the parameter action is 2 and a mailbox does not already exist. This is not how open.message(...) seemed to be working Problem Description (Technical terms) When executing this function with the option 2 there should be a check if the mailbox is already opend, if not the value -1 should be returned, however this is not done, there is only a check if the file is present. Test Procedure execute the next command: function main() { long ret string in.queue(132) in.queue = creat.tmp.file$(bse.tmp.dir$()) ret = open.message(1, in.queue, 2) ret = open.message(1, in.queue, 0) ret = open.message(1, in.queue, 2) if len(in.queue) <= 128 then ret = seq.unlink(in.queue) endif } the ret should be -1 the first, the next time a value should be returned, and the last time the same value should be returned Motive source DF:208185 MaintReger: # 14283 (BDNT7337): BSE_TMP found not accessable on Windows95, 98 or Me Date: Thu, 6 Jun 2002 14:29:22 +0200 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) The following message is logged on Windows95, 98 or Me: Log_mesg: c:\program files\baan\tmp: not accessable; defaulting to TEMP directory Althought the directory c:\program files\baan\tmp is accessable. Problem Description (Technical terms) Function which does the access check used a windows api, using an attribute which is not supported on Windows95, 98 or Me. Workaround No workaround Test Procedure >From a Windows 96 machine: Using BW, log in Baan. Print something to a Windows Printer, having Preview mode switched on. Check your BW.log, or BWPRINT.log file: no message, as described in Problem Desc. should be logged. Also the .bpf file should be created in the BSE_TMP directoy, and not in the TEMP directory. Scopus defectnumber(s) Scopus casenumber(s) Priority (e.g. 10) 123005 666647 20 MaintReger: # 14277 (BDUX7982): display utc date in the debugger Date: Wed, 5 Jun 2002 11:33:26 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) When clicking on the BaanC debugger on a UTC-date variable, it's numeric format is shown. Problem Description (Technical terms) Same as Customer Terms. Workaround Putting the variable name into the debugger's history list, and now the user can display it's value with any kind of format. Test Procedure Click on any variable in the debugger, and it will appears in the history list. Motive source TCS:106-125476 MaintReger: # 14274 (BDUX7911): Bad performance printing to direct printer Date: Tue, 4 Jun 2002 16:45:48 +0200 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) Printer (Mannesman Tally T-661) very slow after installation port. 6.1c.06.03 Problem Description (Technical terms) A reactivate font was done in function hpos() every time a new tab position was set. Reactivating a font is very time consuming on some printers. This reactivation font was introduced to fix a problem with HMI settings after printing Euro signs on HP Laserjet printers (BDUX6286). Workaround n/a Test Procedure print to a direct printer Motive source DF:124929 MaintReger: # 14271 (BDUX7828): Informix:Error on insert after dropping and recreating a table:TCS-100/126499 Date: Tue, 4 Jun 2002 16:43:23 +0200 Created on: DevEscher Type: bugfix Problem Description (Customer terms) None Problem Description (Technical terms) All cursors open on a table are not closed when a table is dropped. The following sequence of operations will cause an error. 1. insert into a table 2. drop the table 3. create the table 4. insert into the table The second insert will reuse the cursor from the first insert and will fail because Informix detects that the table linked to that cursor is dropped. Test Procedure Logon to system Eddie. Go to directory "/home/dbregtr/reg_tests/ctsid108/tcid100/tc4" Run: "bdbpost6.2 -kn -c 623 < pre.dmp" The pre.dmp contains table ddcts550 in company 621 and 622. The "-c" option force that both tables are created in company 623. So table ddcts550 is created in company 623 and the rows are inserted. With a driver that does not include this fix the second post action fails with the error 502. Motive source TCS:100-126499 MaintReger: # 14261 (BDUX7959): Informix:Fixing error 410 on insert:TCS-100/127391 Date: Tue, 4 Jun 2002 16:34:42 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) None Problem Description (Technical terms) Improper freeing of cursors. A cursor that is open in order to insert rows into a non-existent tables is not closed. An attempt to insert rows on the same table of a logical company that refers to the first company causes the cursor from the first insert to be reused and results in an error. Test Procedure Run the following gcommand script #1 ddcts531 TDRTBL 0 -c 621 #1 ddcts531 RINS 0 -i -c 621 -E NO_TABLE = ddcts531.int1 = 1 #1 ddcts531 DABORT 0 #1 ddcts531 RINS 0 -i -c 622 = ddcts531.int1 = 1 (Company 622 has to be a logical company referring to the physical company 621) The driver should return error 506 ( Table does not exist ) for the second insert command. A driver without this fix would return the native error -401. Motive source TCS:100-127391 MaintReger: # 14259 (BDNT7970): Redid change of BDNT7932: was not a problem! Date: Tue, 4 Jun 2002 16:33:44 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) Needed to undo BDNT7932. The problem mentioned in that text was not really a problem, as stated later on by the originator of the call. Motive source TCS:100-125228 MaintReger: # 14239 (BDUX7967): Added cf file for Forte Developer 6 upd compiler for SUN Solaris 2.6 and higher. Date: Mon, 3 Jun 2002 14:13:06 +0200 Created on: MaintReger Type: porting specific change Problem Description (Technical terms) Added cf file for new compiler. However, this file is currently not used, execpt for Open Galaxy. Workaround N/A. Test Procedure For ClearCase systems: don't forget to change Project file in your object tree under Util/imake subdirectory! MaintReger: # 14235 (BDUX7964): Oracle driver should translate ORA-24345 into Baan error 512 Date: Fri, 31 May 2002 14:21:19 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) When a table column in Baan is shorter than the actual size in Oracle, truncation occurs during fetch. Depending on resource settings and Oracle version this can return error ORA-24345. Error ORA-24345 should be returned as E_BDB_DD_CORRUPT (bdb_errno 512) Test Procedure Modify the dtt.pd file. Change the length of ttdsca from 30 to 10 and run the following command: qptool6.2 -c 0 -q "select * from ttaad100" The old driver causes the following output containing the native Oracle error: Error 25345 Msg '' The new driver causes the expected error to occur: Error 512 Msg '' Do not forget to restore the dtt.pd file! Motive source TCS:106-124615 MaintReger: # 14223 (BDUX7963): Force compiler to use ANSI setting, since the default is K&R Date: Fri, 31 May 2002 09:40:12 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Customer terms) Due to compiler settings imake/cpp didn't compile, and therefore no build of the portingset could be started. Problem Description (Technical terms) see above. Workaround Make ANSI the default since old compiler versions take K&R as default (or upgrade compiler) Test Procedure Build! MaintReger: # 14218 (BDUX7961): add include dir Date: Fri, 31 May 2002 08:46:57 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) On W2k sqldmoid.h can not be found by Visual Studio. MaintReger: # 14215 (BDUX7957): lp command fix in the baanC debugger Date: Fri, 31 May 2002 08:44:49 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) in the debugger there is the command L and the command Lp. The command Lp gives less output than the command L. For instance the field "fattr.currfld$" is not printed though it is showed with command L. Problem Description (Technical terms) After the first page dumped into the temporary file, the subfunction cannot open another one to the top of the first one, when the second page dumping is called. Workaround Fixing the used_vars functions. Test Procedure do an lp command in the debugger and check the results with the simple l command's results. Motive source TCS:103-24179 MaintReger: # 14213 (BDNT7956): Add merge6.2 to portingset Date: Fri, 31 May 2002 08:43:10 +0200 Created on: MaintReger Type: bugfix Problem Description (Technical terms) Unix script merge6.2 developed for SCM, ahs to be part of the portingset Test Procedure Run the SCM merge utility MaintReger: # 14201 (BDNT7932): Nested queries: are not placed in busdoc correctly Date: Wed, 29 May 2002 19:44:24 +0200 Created on: MaintCorelli Type: bugfix Motive source TCS:100-125228 MaintReger: # 14197 (BDNT7284): BW5: BW Policies Date: Wed, 29 May 2002 16:22:26 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) The registry key for configuration save at HKEY_CURRENT_USER\SOFTWARE\Policies\Baan\bw is causing problems under Win2K with Normal users. Problem Description (Technical terms) Request for alternative locations to retreive policies from. Workaround Change permissions of current registry keys from which policies (Option1/Option2) are read from. Test Procedure By applying all kind of combinations (set policies in bwcentral.ini file, alternative and normal location in registry), and test. Motive source Scopus call 207428 MaintReger: # 14185 (BDUX7935): Informix:Enabling the Prepare Statement Cache Date: Wed, 29 May 2002 08:46:31 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) None Problem Description (Technical terms) Enabling the prepare statement cache. Test Procedure None MaintReger: # 14182 (BDUX7900): Print with Baan user on AS400 DF123378 Date: Wed, 29 May 2002 08:43:06 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) The problem is that when printing on AS400 as Baan user and the Baan user is not known on the AS400 as Operating System user nothing is printed. E.g. a user (i.e. bsp310) is created and is working on AS400 as user bsp. If a print action is executed to Device Type: printer nothing is printed. Problem Description (Technical terms) The summary of the problem is that the program lp6.2 is called with an option -u and the user behind that option. That useri is the Baan user and not the operating system user. If now the Baan user is different then the operating system user (normal setting for job users) then the lp6.2 command is called with a -u parameter calling an operating system user that does not exist. Workaround Make sure that the Baan user is also known as a OS user. Test Procedure Login to Baan on an AS400 system using a user which is not known as OS user on the AS400. Try to print something to Device Type: printer. Motive source DF:123378 MaintReger: # 14178 (BDUX7934): remove blat API from delivery Date: Wed, 29 May 2002 08:41:06 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) removed blat API drom delivery. We are not givving full API support on it. MaintReger: # 14170 (BDUX7929): Informix:Error 214 in drop table:100-127317 Date: Wed, 29 May 2002 08:34:39 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) None Problem Description (Technical terms) A drop table with the IGNORE REFERENCES flag fails with the Informix error -214 (Cannot remove file for table table-name) Test Procedure Run the following gcommand script 1 ddcts502 TDRTBL -0 1 ddcts502 TDRTBL -1024 The first drop table command will fail due to references. The second one will succedd due to the IGNORE REFERENCES (1024) flag. A driver without this fix would cause the second command to fail as well. Motive source TCS 100-127317 MaintReger: # 14166 (BDNT7917): Bshell: fmax should take number of digits into account Date: Tue, 28 May 2002 14:14:29 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) domain min and max do not take 'Digits before Decimal' en 'Digits after Decimal' into account Test Procedure see TCS call Motive source TCS:106-124658 MaintReger: # 14155 (BDUX7924): Debugger window size bugfix Date: Tue, 28 May 2002 08:46:40 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) The debugger crashes after resizing it's window size under the splitted line. Problem Description (Technical terms) After resizing the debugger it gets smaller value in the total length, then the split line. The bshell gets illegal values, and crashes the debugger process. Workaround Setting default variables if the h-file contains wrong values. Test Procedure start session with debugger, when debugger stops, make command area in debugger larger, by pressing 3 times Split+5 (this is relevant), resize debugger from botton to top in command area, and now it will not crash anymore. Motive source TCS:106-123362 MaintReger: # 14146 (BDUX7925): New bw number and release number due to recent (last) shuttle Date: Mon, 27 May 2002 08:59:06 +0200 Created on: MaintReger Type: porting specific change Problem Description (Technical terms) New version number for 7.3a.02.. Test Procedure check bshell -v MaintReger: # 14139 (BDUX7848): Informix failure encountered on flush cursor Date: Fri, 24 May 2002 13:09:40 +0200 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) The log.informix ($BSE/log) is getting full with the following error : Log_mesg: Informix failure encountered on flush cursor. Problem Description (Technical terms) When using array interface errors that will be handled by the application will be logged in log.informix also. This kind of errors are not real errors because they are expected. e.g. duplicate values during insert records Workaround n/a Motive source DF:124021 MaintReger: # 14132 (BDUX7816): Informix:Map Informix error 217 to Baan error 512:TCS 100-126653 Date: Thu, 23 May 2002 14:12:33 +0200 Created on: DevEscher Type: bugfix Problem Description (Customer terms) None Problem Description (Technical terms) The Informix error 217 (Invalid column name) is not mapped to the Baan error 512 (Corrupt DD) Test Procedure Modify the dd file for a table that already exists in the database by adding a new column. Now run a query on that column. The driver should return error 512. Motive source TCS:100-126653 MaintReger: # 14126 (BDUX7884): bdbpost skips table on error. Date: Thu, 23 May 2002 11:59:24 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) bdbpost6.x skips loading next table. Problem Description (Technical terms) If multiple tables are loaded with bdbpost with the "-f" option and the creation of an index fails, the next table to be loaded will be skipped and bdbpost continues with the second next table. Workaround No workaround. Test Procedure Create with bdbpre6.2 an export file of three tables and drop the tables. For example: $ bdbpre6.2 -k -N dbhvl010 dbhvl011 dbhvl012 -C 700 > table.dmp Create in the "storage" file of the used database a line for the indexes of the first table with a not existing indexspace. For example: (Oracle): dbhvl010:611:I:TABLESPACE exist_not PCTFREE 5 INITRANS 3 Recreate the table with bdbpost from the export file. For example: $ bdbpost6.2 -knf < table.dmp The output of the old and wrong bdbpost likes: Default separator ('\0') taken Table M Idx RowCount Messages ----------- - --- -------- ---------------------- dbhvl010700 1 3 Adding indexes ...Can not creat index 1 (error 1959) mkdir error 22 dbhvl012700 1 6 Adding indexes ... The output of qptool shows that the second table isn't created. For example: $ qptool6.2 -q "select * from dbhvl011" -c 700 [0] SQLState QP999: Table does not exist (error 506) The output of the new and correct bdbpost likes: Default separator ('\0') taken Table M Idx RowCount Messages ----------- - --- -------- ---------------------- dbhvl010611 1 3 Adding indexes ...Can not creat index 1 (error 1959) dbhvl011611 1 3 Adding indexes ... dbhvl012611 1 6 Adding indexes ... The output of qptool shows that the second table is created. For example: $ qptool6.2 -q "select * from dbhvl011" -c 700 1 730570 985608869 'One ' 700 2 730571 985608876 'Two ' 700 3 730572 985439687 'Three ' 700 In both cases the exit code of bdbpost is unequal to zero because the index(es) of the first table can't be created. So the first table has no indexes after the import with bdbpost. Motive source DF:113869 MaintReger: # 14125 (BDUX7823): DB2:Map DB2 error 206 to Baan error 512:TCS 100-126655 Date: Thu, 23 May 2002 11:00:52 +0200 Created on: DevEscher Type: bugfix Problem Description (Customer terms) None Problem Description (Technical terms) The DB2 error -206 (Invalid column) is not mapped to the Baan error 512 (Corrupt DD) Test Procedure Modify the dd file for a table that already exists in the database by adding a new column. Now run a query on that column. The driver should return error 512. Motive source TCS:100-126655 MaintReger: # 14120 (BDNT7880): MSQL_SERVERHOST not taken in account by msql7_maint.exe Date: Thu, 23 May 2002 10:30:40 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) When running msql7_maint.exe in batch mode, the following is specified in the command file. E.g.: -F -OMSQL_LEVEL1=1,MSQL_SERVERHOST=SRV-NLD-BCI0 -F -abaanadm -mbaanadm -p)12I,M+Lq5%vb`#xI,bj|Xj\O"4u"q.Q -csa -iHB$BAkD?iE"^T]-vg2w5M&PgZ<^pr~;~ -Gbaandb -PpDOH9u~?Tkv~*3e!&t#