Release notes for portingset: 7.1d.02 MaintCorelli: # 18001 (BDUX11432): Fixed compilation errors on HPUX10 Date: Tue, 15 Apr 2003 12:23:04 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) Fixed compilation error on HPUX10 for C++ stuff (idlcompiler) Test Procedure Build idlcompiler (debug build) MaintCorelli: # 17998 (BDUX11427): build error OS390 Date: Tue, 15 Apr 2003 11:21:44 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) nightly build failed MaintCorelli: # 17990 (BDUX11086): Return value of bdbpost in case of drop of logical table Date: Mon, 14 Apr 2003 12:50:03 +0200 Created on: MaintReger Type: bugfix Problem Description (Customer terms) 1. When importing dumps having "fixed length" and "seperator" (*.S files), only the last imported table determines the returnvalue of bdbpost. 2. When bdbpost tries to drop a logical table, this is interpreted as a critical error by the 4GL, because bdbpost returns 10 in that case. But because it concerns a logical table this error can be ignored. (the drop table is not done on a logical table, and because the drop is not done, also no data is imported). So the 4GL should not generate an error in this case. Problem Description (Technical terms) 1. After a table is imported not-successfully the ret_val should not be reset by a table which is imported successfully. 2. When bdbpost tries to drop a logical table, it returns 10, and does not do an import of the data. Instead of returning 10 it should return 0 (and not do an import of data). Workaround Not available. Test Procedure The following company mappings is applied (compnr6.2): tccom000:200:998 tcemm:200:998 so table tccom000 of logical company 998 is mapped to physical company 200 and all tables of modulethat is no critical errors. That's ok. The log.bdbpost6.2 contains the following errormessages for tables which could not be dropped: 2003-03-18[09:23:57(UTC+05:00)]:E:fvdveer: *******cated at: /vobs/tt/lib/sql/Logger/SQLLogger.cpp:68, size=16 5.2> echo $? 7 Now the returnvalue of bdbpost is 7 (Can not read table definition). That is correct, because there was 1 table giving error 513. Motive source DF:134269 MaintCorelli: # 17966 (HYNT11402): MSQL: Password-protect the L1->L2 fast migration tool Date: Fri, 11 Apr 2003 14:01:15 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) There is no password protection for the migration tool. It is necessary to prevent it falling into 'uncontrolled usage situation' (business sense;-). Test Procedure When you run the migration tool, it should ask for a password and should proceed only on validating it. MaintCorelli: # 17965 (HYNT11092): MSQL: NOIDXHINT option enabled in msql_driver_param file Date: Fri, 11 Apr 2003 14:00:25 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) NOIDXHINT cannot be specified in msql_driver_param file, which is the right place for this option. Workaround NOIDXHINT can be specified in (the legacy) msql_storage file Test Procedure Specify the NOIDXHINT in the msql_driver_param in one or many entries (at the end of each entry). Test it for a single table, more than one table with/without specifying the NOIDXHINT. Check the actual query generated by the driver (in the dbs.log). The Index hint should not be present when NOIDXHINT is specified. MaintCorelli: # 17947 (BDUX11389): Fixed compilation error on Linux (and may be others) Date: Thu, 10 Apr 2003 15:04:11 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) Fixed build problem on Linux Test Procedure run Util/Boot on Linux. Check that no error occurs MaintCorelli: # 17941 (BDUX11388): Fixed compilation errors on Siemens Date: Thu, 10 Apr 2003 14:36:17 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) Siemens shared libs were not build due to (bug in ?) compiler disabled pic code with -FX4y flag. Test Procedure Build jvm2bvm shared library MaintCorelli: # 17940 (BDUX11231): restore hpux11_pa2.0.cf Date: Thu, 10 Apr 2003 14:04:37 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Less performance HPUX11 portingset. Problem Description (Technical terms) Since this portingset is build with DAportable iso DA2.0 there was confusion about the Machine-id: HP9000_PA_RISC . It is changed back to HP9000_PA_8000 . In the latest builds no performance improvement was done by means of chatr command, from now on this will be done again. MaintCorelli: # 17928 (BDUX11345): Cant find type XXX in LC_REGISTER errors Date: Thu, 10 Apr 2003 11:37:11 +0200 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) In a network license in Baan IV and ERP customer is getting the following error messages: In Baan IVc Log_mesg: set_value: Can't find type 399 in LC_REGISTER Log_mesg: set_value: Can't find type 653 in LC_REGISTER Log_mesg: set_value: Can't find type 654 in LC_REGISTER In BaanERP Message: set_value: Can't find type 399 in LC_OK Message: set_value: Can't find type 653 in LC_OK Problem Description (Technical terms) This problem is only applicable in case of network licence, so there is a "oserver." At the moment the command LC_USERS is created first the values are set, and then the set_command is done. This is not correct. First the set_command should be done, and then the set_values for this command. Because the set_command is done after the set_values, the set_value function generate these error, because it does not except this field types. (The header of the previous command is still in the memory). Workaround No workaround available. Test Procedure Have at least 2 Baan servers, validated for a Network licence. Have at both servers the licence daemon running. At both servers execute the command: licmon6.1 -u Then login at both servers into Baan a couple of times. You can make a .bat file to do this in an automated way. And log out again. After logging out, no errors should be logged in the log.licd6.1 file. Motive source DF:213393 MaintCorelli: # 17918 (BDUX11379): Typo in Imake.rules Date: Wed, 9 Apr 2003 14:52:44 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) Typo in rule. Replaced comma by dot MaintCorelli: # 17912 (BDUX11193): Merged project BDUX10917 to Corelli (also use right compilation flags for shared libraries) Date: Wed, 9 Apr 2003 13:46:32 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) Let shared libraries use the regular compilation flags for optimizing/debug Note: an additional fix is required for Siemens (will follow) Test Procedure Build a shared library (e.g. jvm2bvm) and check that the compilation flags contain now the same flags as defined in the COptimizedDebugFlags define from configuration (.cf) file from ConfigTools/Util/imake directory MaintCorelli: # 17887 (BDNT11170): Slow startup time of first session on Windows Date: Tue, 8 Apr 2003 16:03:37 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Startup of session is slow, in spite of the fact that the tabledefinition files for that package combination are pre-loaded into the Baan Shared Memory. Problem only applicable for Windows platform. Specially sessions, which needs a lot of tabledefinitions, are very slow. Because the tabledefinitions are pre-loaded into the Baan Shared Memory you should expect that also first time startup is fast. For example: When you start session tdsls4101m000 for the first time, it takes about 15 seconds before the session is completely started and available. Problem Description (Technical terms) In case the bshell needs to obtain the tabledefinition info for the first time, an explode is done to determine the pathname. E.g.: d:\baan\tools\dd\dttaad\dttaad200. This explode reads the search path information from the fd file, to determine the location of the tabledefinition. During this explode action, the physical file on disk is also opened to determine whether the file really exists. And this open action consumes a lot of time, specially on Windows. When a session needs a lot of tabledefinitions, the total time of all these open actions can become big: 10-20 seconds. In case the file does also exists on disk, this pathname is taken to search in SHM. When found in SHM, the contents of the file itself are read from Shared Memory (SHM). (Unless the file is not in SHM, in that case the contents will still be read from disk) . Workaround There is no workaround for this issue. Test Procedure On a Windows platform: 1. Take care that all tabledefinitions are pre-loaded into Baan Shared Memory for the package combination you are working in. 2. Fill in the following in the BW Command field: -- -dbgsrdduse -dbgfdev -logfile trace_shm -keeplog tdsls4101m000 3. Start BW, when session tdsls4101m000 is started, close the session window again and log out of Baan 4. View the file trace_shm (created in BSE_TMP directory). 5. It should show you for example: _explode :(d:/baan4c/tools/dd/dttgfd/dttgfd303) explode: F_BRDD:dttgfd303, d:/baan4c/tools/dd/dttgfd/dttgfd303, d:/baan4c/tools/dd/dttgfd/dttgfd303 get_explode_fname(d:/baan4c/tools/dd/dttgfd/dttgfd303) Table dump d:/baan4c/tools/dd/dttgfd/dttgfd303 load from srdd (no disk check done) And there should be no "sopen" for the tabledefinitions, which are in Shared Memory. 6. Now do the same test, but then take care that the all the tabledefinitions are NOT pre-loaded into the Baan Shared Memory. 7. View the trace file, and it should show you now for example: _explode :(d:/baan4c/tools/dd/dttgfd/dttgfd303) explode: F_BRDD:dttgfd303, d:/baan4c/tools/dd/dttgfd/dttgfd303, d:/baan4c/tools/dd/dttgfd/dttgfd303 get_explode_fname(d:/baan4c/tools/dd/dttgfd/dttgfd303) Table dump d:/baan4c/tools/dd/dttgfd/dttgfd303 load from disk _explode :(d:/baan4c/tools/dd/dttgfd/dttgfd303) explode: F_BRDD:dttgfd303, d:/baan4c/tools/dd/dttgfd/dttgfd303, d:/baan4c/tools/dd/dttgfd/dttgfd303 open(d:/baan4c/tools/dd/dttgfd/dttgfd303), fd 5 errno 0 sopen(d:/baan4c/tools/dd/dttgfd/dttgfd303) 0x52f388 Now, there are "sopen" for the tabledef's which are loaded from disk. Beside this test procedure you can also test using an old bshell (not containing this fix), and a new one. My test showed out, when using the old bshell, the startup time was 17 seconds, and when using the new one, the startup time was about 2 seconds. Motive source DF:213234 MaintCorelli: # 17873 (BDUX11336): Fixed coredump on db_read_eq Date: Mon, 7 Apr 2003 12:21:08 +0200 Created on: MaintBaanIVc Type: bugfix Problem Description (Technical terms) Coredump on db.read.eq in Level-2 database drivers MaintCorelli: # 17868 (BDUX11331): Build fix for non AIX systems Date: Mon, 7 Apr 2003 12:16:52 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) Last compilation fix were work only for AIX systems. The non-AIX systems still cannot compile the bshell. Test Procedure Build bshell6.2 MaintCorelli: # 17855 (BDUX11324): Build failed on Corelli Date: Fri, 4 Apr 2003 12:22:28 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) Nightly build failed for Corelli caused by a wrong fix. Test Procedure Compile bshell. MaintCorelli: # 17850 (BDUX11323): AS400: Add ibm_as400V5R2.cf file Date: Fri, 4 Apr 2003 11:05:36 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) OS400 version V5R1 will be EOL by 31 may 2003. MaintCorelli: # 17847 (BDUX11322): AS400: Add ferret as build system Date: Fri, 4 Apr 2003 11:03:09 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) Ferret is an AS400 V5R2 Logical Partion for builds. MaintCorelli: # 17846 (BDUX11267): Database driver loops when _storage_param file is empty Date: Fri, 4 Apr 2003 11:02:19 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Database driver loops when _storage_param file is empty. e.g. create an empty _storage_param file in $BSE/lib/ perform a table action (create table) driver takes now 100 % CPU. Problem Description (Technical terms) Function qdb_get_storage_param() will not leaved in case EOF was reached. Workaround A matching entry should be in _storage_file. Motive source DF:135101 MaintCorelli: # 17839 (BDUX11257): Divzero fix Date: Thu, 3 Apr 2003 16:29:26 +0200 Created on: DevNewton Type: bugfix Problem Description (Technical terms) Bshell failed to start because it detects division by zero event. This was caused a fix where a common signal were used to detect this event. Unfortunately JavaVM is also using that signal for internal exception handling. Test Procedure Execute Webtop. MaintCorelli: # 17821 (BDUX11247): DB drivers: sync array fetching with AS SET WITH ROWS Date: Thu, 3 Apr 2003 13:35:47 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Performance issue when array fetching has been enabled for queries containing AS [PREPARED] SET WITH ROWS clause. Problem Description (Technical terms) The following AS SET WITH 1 ROWS query fetched too much records from Oracle because array fetching was enabled and set to a high value: select tdsls312.rlno, tdsls312.rlrv, tdsls312.rtyp, tdsls312.sodb, tdsls312.stbp from tdsls312 where tdsls312._index4 = {:hold.ofbp, :g.customer.rel.no} and tdsls312.rtyp = :ecedi005.rtyp as set with 1 rows selectdo endselect resulteerd in (TKPROFF): SELECT /*+ FIRST_ROWS index(a ttdsls312605$idx3) */ a.t$rlno,a.t$rlrv, a.t$rtyp,a.t$sodb,a.t$stbp FROM baan.ttdsls312605 a WHERE a.t$rtyp = :1 AND a.t$ofbp = :2 AND a.t$crln = :3 call count cpu elapsed disk query current rows ------- ------ -------- ---------- ---------- ---------- ---------- ---------- Parse 448 0.42 0.31 0 0 0 0 Execute 448 0.07 0.08 0 0 0 0 Fetch 448 305.16 326.05 0 34758080 0 13821 ------- ------ -------- ---------- ---------- ---------- ---------- ---------- total 1344 305.65 326.44 0 34758080 0 13821 Only 448 rows should have been fetched from Oracle instead of 13821, eliminating a performance issue. Workaround Be carefull with high array fetch settings Test Procedure Set Environment: RDS_FULL=10 ORA_INIT=0111000 ORA_MAX_ARRAY_FETCH=100 DBSLOG=01570 Run: qptool6.2 -c 0 -q "select * from ttadv999 as set with 3 rows" Dbs.log of the old driver contains: rds_full E 10 ora_max_array_fetch E 100 SQL> SELECT /*+ FIRST_ROWS */ a.t$desc,a.t$keyr,a.t$pacc FROM b52.tttadv999000 a SQL> ora_parse( 0x400bcf40 ) SQL> ora_multi_execute( 0x400bcf40 ) do 0, prefetch 1 SQL> ora_multi_fetch( 0x400bcf40 ) no_rows 100 <<< picked up array size Dbs.log of the new driver contains: rds_full E 10 ora_max_array_fetch E 100 Array size exceeds WITH ROWS: reduced to 3 SQL> SELECT /*+ FIRST_ROWS */ a.t$desc,a.t$keyr,a.t$pacc FROM b52.tttadv999000 a SQL> ora_parse( 0x400bcf40 ) SQL> ora_multi_execute( 0x400bcf40 ) do 0, prefetch 1 SQL> ora_multi_fetch( 0x400bcf40 ) no_rows 3 <<< Reduced array fetching to Run: qptool6.2 -apd40ef*bc -c 0 -q "select * from ttadv999 as set with 3 rows" Dbs.log of the old driver contains: rds_full E 10 ora_max_array_fetch E 100 SQL> SELECT /*+ FIRST_ROWS */ a.t$desc,a.t$keyr,a.t$pacc FROM b52.tttadv999000 a SQL> ora_parse( 0x400bcf40 ) SQL> ora_multi_execute( 0x400bcf40 ) do 0, prefetch 1 SQL> ora_multi_fetch( 0x400bcf40 ) no_rows 40 <<< picked up rds_full set to 40 SQL> ora_cancel_cursor( 0x400bcf40 ) Dbs.log of the new driver contains: rds_full E 10 ora_max_array_fetch E D 100 Array size exceeds WITH ROWS: reduced to 3 SQL> SELECT /*+ FIRST_ROWS */ a.t$desc,a.t$keyr,a.t$pacc FROM b52.tttadv999000 a SQL> ora_parse( 0x400bcf40 ) SQL> ora_multi_execute( 0x400bcf40 ) do 0, prefetch 1 SQL> ora_multi_fetch( 0x400bcf40 ) no_rows 3 <<< Reduced array fetching to SQL> ora_cancel_cursor( 0x400bcf40 ) Run: qptool6.2 -apd2ef*bc -c 0 -q "select * from ttadv999 as set with 3 rows" Dbs.log of the old driver contains: rds_full E 10 ora_max_array_fetch E 100 SQL> SELECT /*+ FIRST_ROWS */ a.t$desc,a.t$keyr,a.t$pacc FROM b52.tttadv999000 a SQL> ora_parse( 0x400bcf40 ) SQL> ora_multi_execute( 0x400bcf40 ) do 0, prefetch 1 SQL> ora_multi_fetch( 0x400bcf40 ) no_rows 2 <<< picked up rds_full set to 2 SQL> ora_multi_fetch( 0x400bcf40 ) no_rows 2 SQL> ora_cancel_cursor( 0x400bcf40 ) Dbs.log of the new driver contains: rds_full E 10 ora_max_array_fetch E 100 SQL> SELECT /*+ FIRST_ROWS */ a.t$desc,a.t$keyr,a.t$pacc FROM b52.tttadv999000 a SQL> ora_parse( 0x400bcf40 ) SQL> ora_multi_execute( 0x400bcf40 ) do 0, prefetch 1 SQL> ora_multi_fetch( 0x400bcf40 ) no_rows 2 <<< picked up rds_full set to 2 SQL> ora_multi_fetch( 0x400bcf40 ) no_rows 2 ||| No reduction as max_rows > rds_full SQL> ora_cancel_cursor( 0x400bcf40 ) MaintCorelli: # 17816 (BDUX11220): db.check.row.changed updates reference counters Date: Thu, 3 Apr 2003 11:45:26 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) Currently a check for row change(invoking db.check.row.changed) also generates update on reference counters thus corrupting the counters in the database. Test Procedure Though this can be tested in numerous ways, the simplest would be to write a small 3GL script and invoke it using ba6.2. The script could be something like the following: long __fp, __ret #define PUTS( S ) ^ if ( __fp = 0 ) ^ then ^ __fp = seq.open( getenv$("SQL_OUTPUTFILE"),"a" ) ^ endif ^ __ret = seq.puts( S, __fp ) #define PUTF( S, ... ) ^ PUTS( sprintf$( S, ... ) ) #define ASSERT( X ) ^ if ( not ( X ) ) ^ then ^ PUTF( "ASSERT failed; line=%d, file=%s", __LINE__, __FILE__ ) ^ abort() ^ endif table tdbtst120 function main () { long ret db.first(tdbtst120, db.delayed.lock) dbtst120.workdept = "E01" ret = db.check.row.changed(tdbtst120) ASSERT( ret = 0 ) } Execute this script with the options DBSLOG=1570 and verify that the "dbs.log" DOES NOT contain the following information to update the reference counters which was the case prior to the fix. ... SQL> UPDATE reger.tdbtst120000 a SET t_Refcntd=t_Refcntd+ ?,t_Refcntu= t_Refcntu+ ? WHERE t_workdept= ? qdb_open_update_cursor done. ... Motive source IRD: 38 MaintCorelli: # 17796 (BDUX11275): Minor bug in cd_read, occurring on Linux Date: Wed, 2 Apr 2003 15:42:39 +0200 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) cd_read on Linux used less program to show content of files. Less caused confusion, because 'q' need to be pressed, when shown file reached EOF Workaround type 'q' when less has shown the installation info (readme.mas) Test Procedure Run cd_read6.x and verify that no 'q' key press is needed anymore MaintCorelli: # 17792 (BDUX11231): restore hpux11_pa2.0.cf Date: Wed, 2 Apr 2003 12:42:51 +0200 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Less performance HPUX11 portingset. Problem Description (Technical terms) Since this portingset is build with DAportable iso DA2.0 there was confusion about the Machine-id: HP9000_PA_RISC . It is changed back to HP9000_PA_8000 . In the latest builds no performance improvement was done by means of chatr command, from now on this will be done again. MaintCorelli: # 17782 (BDUX11202): undo BDUX10693 Date: Tue, 1 Apr 2003 17:09:44 +0200 Created on: DevEscher Type: porting specific change Problem Description (Technical terms) Floating Point Exeption on AIX when JVMI is used. The SIGTRAP is also used by jvm. Signals are not handled correctly. Test Procedure use WebTop MaintCorelli: # 17756 (BDUX11216): When wrong cf file is in object tree, Boot should halt Date: Fri, 28 Mar 2003 16:13:16 +0100 Created on: DevEscher Type: porting specific change Problem Description (Technical terms) Util/Boot was not able to handle single quotes in CF_CHECK_COMMAND (like used on SUN) This problem only happens currently for all solaris*.cf files MaintCorelli: # 17752 (BDUX11213): Fixed build for Oracle7 driver on AIX Date: Fri, 28 Mar 2003 16:09:37 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) Nightly build fauiled on the AIX build machine due to an error in the Imakefile MaintCorelli: # 17751 (BDNT11212): Manual merge project BDNT11148 to 4c (enhanced logging of SHM) Date: Fri, 28 Mar 2003 16:06:19 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Technical terms) Manual merge of project BDNT11148, enhanced logging for Shared memory Manager/Bshell Workaround NA MaintCorelli: # 17739 (BDUX11207): Oracle resources not set correctly when db_resource file is empty Date: Fri, 28 Mar 2003 12:28:31 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Technical terms) "Index out of dims" error on corrupt screen dumps. This was caused because Oracle translated the data, corrupting the stop byte. This was all caused because during installation the oracle driver resources were not set correctly. Workaround Before installation, create a $BSE/lib/defaults/db_resource file with at least one (empty) line. Motive source TCS:100-131015 MaintCorelli: # 17704 (BDUX11010): Oracle driver: enable use of OCI7 interface Date: Tue, 25 Mar 2003 14:13:01 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) There is a performance gap between the ora7_srv and ora8_srv, especially when a lot of data is retrieved (batch sessions). Problem Description (Technical terms) When array fetching has been enabled, oci7 performs much better (> 20%) than oci8. Test Procedure All available regression tests. Motive source DF-213114 MaintCorelli: # 17701 (BDUX11097): licd6.2 does not write logfile on As400 systems Date: Tue, 25 Mar 2003 12:00:12 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) AS400 portingset : Licence-daemon does not write log-file Problem Description (Technical terms) Logfile was not used in child process of licd6.2. Workaround n/a Test Procedure start licd6.2 and check the logfile in $BSE/log directory. Motive source DF:133354 MaintCorelli: # 17697 (BDNT11073): BW5: Crashes when printing to bwprint Date: Tue, 25 Mar 2003 11:26:50 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) You will be disconnected from the server if you print to a physical printer using the session tisfc0408m000, Worktop 2.2.25, 2.2.26 and Windows98. (Worktop is used centrally deployed, so a shortcut is started on the Windows98 client. This shortcut start the worktop_client from a networkpath, and has as argument a worktop document, also from a networkpath. E.g.: \\cnl01234\worktop\bin\worktop_client.exe \\cnl01234\worktop\documents\test.wtd.) Problem Description (Technical terms) The access violation is caused, because BW tries to update an object, which is already destroyed (object_entry->object is 0). (DmUpdateObjectEntry(object_entry,widget,widget) at #625 dn_wobj.c) This is because the object (it concernes type 48, type DsCapplication), is destroyed to early. The following happens: Bwprint is started by app_start() from the 4gl code. BW handles this functionality. (applet.cpp: TsApplet::start()), and an application object is created. During the creation of this application object (dm_wobj.c: object_entry = DmCreateNewObjectEntry()), new Window events are received on the socket of BW (WS_SELECT_EVENT/FD_READ). These events are handled in the meanwhile by another thread by function processIncomingData, and causes execution of the callback function DmDestroyWinObject. To solve this problem, BW should not process incoming data on the socket as long as the creation of the object is not finished. So, in other words: it should not re-enter the function processIncomingData, as long as it as already processing incoming data, in order to avoid messing up of different messages and message loops. Workaround No workaround. Test Procedure Refer to description of defect 133760. Not possible to define generic test procedure. Motive source DF:133760 MaintCorelli: # 17667 (BDUX11137): Fixed nightly build on NT Date: Mon, 24 Mar 2003 10:28:24 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) Nightly build failed on NT, as a include file could not be found. Test Procedure nightly build MaintCorelli: # 17666 (BDUX11136): Fixed shared libary build on AIX Date: Mon, 24 Mar 2003 10:27:23 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) Nightly build failed on AIX ( dbapi.so ): ld: 0711-317 ERROR: Undefined symbol: .blat_proc_new_server Test Procedure nightly build MaintCorelli: # 17659 (BDUX11128): BLAT: resources for online configuration (port) Date: Fri, 21 Mar 2003 13:45:37 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Several requirements for BLAT - Configuration Coupling with the session: ttstpbshdebug Coupling of bshell and database driver traces (configuration changes in the bshell are propagated to the database drivers). - Extra logging PROC_TEXT user remarks to the trace (kind of "I was here" tag) PROC_STATE log trace state changes Problem Description (Technical terms) The $BLAT environment variable has been upgraded to a resource. This means that it can be changed using the toggle.cpu() bshell function. The session ttstpbshdebug is using this. Workaround Keep on using the current complex enabling MaintCorelli: # 17649 (BDNT11109): Bwprint: log usefule error in case barcode image cannot be generated Date: Thu, 20 Mar 2003 16:24:21 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) When a barcode is not specified correctly, e.g. wrong barcode length, bwprint just does not show or print the barcode image. It is a hard time to figure out what the reason is. Therefore a useful error would be nice. Problem Description (Technical terms) No error generated in case barcode image cannot created. Workaround Check out the contents of the bpf file for the control codes representing the barcode. Test Procedure Specify a barcode in a report which specifications are not according the "rules" of that barcode. And print this report to a Windows printer device. Bwprint should log an error in that case. You can also do the Barcode test in bwprint, and save this test as another file, modify the file manually, so it contains wrong specified barcodes. Motive source DF:135037 MaintCorelli: # 17639 (BDUX10842): asterisk fill on numeric fields Date: Thu, 20 Mar 2003 10:46:03 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) When using the format "*$ZZTZZTZZ9VD99" to print numeric amounts on checks and having a asterisk fill to prevent tampering the following happens for various numbers: 189.57 - $****** 189.57 - wrong, leaves blank in front of "1" 3163.00 - $*****3,163.00 - good, completely filled with stars 98765.00 - $*** 98,765.00 - wrong, leaves blank in front of "9" 1127815.07 -$*1,127,815.07 - good, filled with stars appropriately It appears that if there are an odd number of digits, one blank is put in front of the amount. If there an an even number of digits, it looks good. Problem Description (Technical terms) The function str_edit calls function 'vullen', which is responsible for filling spaces or asterisks in a string representing a numeric value. This is all depending on the defined format. In the function 'vullen' a hardcoded space is entered in the string in some situation. This is wrong because in case the value has to be printed on a cheque there cannot be any spaces. Whether it must be a space or an asterisk is given to the function 'vullen' as an argument (vulsel). Instead of a space the value 'vulsel' is now used. Workaround N/A Test Procedure The best way to test this fix is to create a little 3GL script like: function main() { double rjr string my(32) rjr = 123 my = edit$(rjr, "*$ZZTZZTZZ9VDZZ") rjr = 123.45 my = edit$(rjr, "*$ZZTZZTZZ9VDZZ") rjr = -123.45 my = edit$(rjr, "*$ZZTZZTZZ9VDZZ") rjr = 1234.56 my = edit$(rjr, "*$ZZTZZTZZ9VDZZ") rjr = 123 my = edit$(rjr, "*$ZZTZZTZZ9VDZZ-") rjr = -123 my = edit$(rjr, "*$ZZTZZTZZ9VDZZ-") rjr = 123.45 my = edit$(rjr, "*$ZZTZZTZZ9VDZZ-") rjr = -3.45 my = edit$(rjr, "*$ZZTZZTZZ9VDZZ-") } This are just some examples, you can use any value or format. Important is the * as first character in the format. This indicates the result must have asterisks instead of spaces. Motive source DF:212775 MaintCorelli: # 17624 (BDUX11107): Merge Enh. BDUX10545 Date: Wed, 19 Mar 2003 11:23:09 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) No timing information available when testing with art functions with use of art trace logging. Problem Description (Technical terms) idem Workaround Do use tracing and ARM products like HP Openview and Tivoli Test Procedure Enable ART in $BSE/lib/defaults/all with art_enable:1 art_trace:1 art_user:baan Start some 4GL session and examine $BSE/tmp/bshell. Motive source DF:212459 MaintCorelli: # 17621 (BDUX11104): AS400: add mink and skunk to patch.c Date: Wed, 19 Mar 2003 10:41:13 +0100 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) For Corellii and Reger, skunk and mink need to run patch6.2 when building a portingset Test Procedure $ vi /vobs/logic/licence/patch.c # add hostid skunk and mink > cd /home/port/MaintCorelli/licence > rm /home/port/MaintCorelli/bin/patch6.2 > gmake /home/port/MaintCorelli/bin/patch6.2 > system -v "CRTSAVF (PORT/PATCH6.2)" > system -v "CLRSAVF (PORT/PATCH6.2)" > system -v "SAVOBJ OBJ(PATCH6.2) LIB(MAINTCOR) DEV(*SAVF) OBJTYPE(*PGM) SAVF(PORT/SCRYPT6.2)" > system -v "RSTOBJ OBJ(*ALL) SAVLIB(MAINTCOR) DEV(*SAVF) SAVF(PORT/SCRYPT6.2) MBROPT(*all) ALWOBJDIF(*ALL) RSTLIB(PORT)" > cd /home/port/MaintCorelli > gmake deliver 2>&1 | tee deliver.out2 MaintCorelli: # 17616 (BDUX11100): Fixed mistake, use right config file for Solaris 2.6 Date: Tue, 18 Mar 2003 16:54:40 +0100 Created on: MaintBaanIVc Type: porting specific change Problem Description (Technical terms) Use right cf file for optimized build on Solaris 2.6 with Workshop 6 upd 2 compiler Test Procedure Build on Solaris 2.6 (e.g. orbix) MaintCorelli: # 17611 (BDUX11091): AS400 Installer: Database resource variable settings Date: Tue, 18 Mar 2003 10:53:25 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Tables are created in Level2 Mode regardless of choice of Level1 Problem Description (Technical terms) In db_resource file db4_db_version:2 entry is added regardless of user choice. Workaround Before import, remove the entry db4_db_version:2 from db_resource file. Test Procedure Try installation with new media. Motive source TCS:100-131234 MaintCorelli: # 17603 (BDUX11083): AS400: Add skunk and mink to build systems Date: Mon, 17 Mar 2003 16:37:21 +0100 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) Programs patch6.2 and scrypt6.2 use the hostid6.2 number to check if files may be scrypted or patched on the build machines. AS400 skunk and mink will replace frite and dilbert. Test Procedure Nightly build MaintCorelli: # 17587 (BDNT11017): Bwprint: add option to use default printer automatically Date: Fri, 14 Mar 2003 14:47:44 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) When running nightly jobs (in the background), sometimes it can happen that a job does not continue to run, because bwprint cannot find the specified printer. At that moment bwprint pops up the message: Printer '' not found, print to default printer?. Because the printsession is running in a job this question is not being be answered. It should be possible to specify that bwprint will use automatically the default windows printer, without asking. Problem Description (Technical terms) Bwprint pops up a message which makes only sense in interactive mode. Workaround Do not run bwprint.exe in the background, or take care that the specified printer is available. Test Procedure The following device is defined in Baan (ttaad3100m000 Device Data): Device Type: Windows Printer Device Queue: \\cnl08696\blablabla Paper Type: A4 Page Length: 66 On my local pc, cnl08696 device "blablabla" does not exists. Print from a voluntary printsession to this device. (Do not check "Preview" option). When printing from Baan to this device, by default behaviour bwprint asks me: Printer '\\cnl08696\blablabla' not found, print to default printer? After pressing OK, the default printer is used. Now, start bwprint (by doubleclicking bwprint.exe), and select Fonttest from the Help. Right click on some text and select "Default Printer" from the drop down menu. Then select "Enable automatic use of default printer" Print again from a voluntary print session from Baan, using the same device. Now, bwprint should not ask you to print to the default printer, but should automatically use the default printer. Now, start bwprint (by doubleclicking bwprint.exe), and select Fonttest from the Help. Right click on some text and select "Default Printer" from the drop down menu. Then select "Disable automatic use of default printer (default)" Print again from a voluntary print session from Baan, using the same device. Now, bwprint should ask you to print to the default printer, and should not automatically use the default printer. Motive source DF:133429 MaintCorelli: # 17581 (BDNT10682): Uppercase conversion of Thai characters went wrong Date: Fri, 14 Mar 2003 14:26:54 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) >From session tcibd0501m000 (General Item Data) session tcibd0101s000 is started. In this session you can define an Item. The program generates 2 search keys: Search Key I, and Search Key II. These search keys are created using an uppercase conversion of the description of the Item. But in case the description contains Thai characters, the Search Key contains wrong Thai characters. Problem Description (Technical terms) The conversion from to Uppercase is done using the standard conversion using the iso8859_upper[] table. But in case of Thai characters this conversion does result in a quite other character, and that is not correct because the Thai characters do not have a seperate lower- and uppercase variant. >From init_iso it is possible to put another table from the file $BSE/lib/nlsinf/, over the iso8859_upper[] table. Therefore the file ISO_BINB is modified to define 1:1 conversion for Thai characters. Workaround Do not use Thai characters in the description of the Item. Test Procedure Have the ISO_BINB file installed in $BSE/lib/nlsinf, and check using session tcibd0101m000 if the Search Key's generated, are correctly when usig Thai characters in the Item description. Motive source DF:132439 MaintCorelli: # 17578 (BDNT10908): Access violation error in bw.exe Date: Fri, 14 Mar 2003 14:24:25 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) The GUI will terminate if a huge layout is openend in reporteditor and you click with the mouse at the bottom of the report. The bw will generate an access violation error and closes. Problem Description (Technical terms) In case the mouse button is released in a none window area and the window where the button was pressed is deleted already the GUI will terminate. Reason for this is that a pointer is used that points to a window which does not exist anymore. Workaround n/a Motive source DF:134146 MaintCorelli: # 17576 (BDUX10913): Ported solaris 8 Date: Fri, 14 Mar 2003 10:16:28 +0100 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) Ported solaris 8 on Corlelli Test Procedure Compile code on Solaris 8 box with Forte 7 compiler MaintCorelli: # 17575 (BDNT11034): Control logging of get_format_entry errors Date: Thu, 13 Mar 2003 16:21:41 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) The error logs are filled up with a lot of get_format_entry error messages. This is request for that the get_format_entry errors are not to be logged when at least a generic format is found for the format code. If the format code & unit combination is not found (eg. Q13/FT but the generic format code does exist (eg Q13) then no error message should be written to the log.bshell file? This would help us tremendously in many ways: 1 - being our log files will not roll over so fast that we miss real error messages. 2 - performance will improve for batch jobs which end up writing thousands of these messages. 3 - we will not need to continously update the datecurr file. Problem Description (Technical terms) Function get_format_entry searches datecurr file for the format specifier. For example: C|001USD|USD|ZZZZ9| Here: C: Means this line specifies a currency format 001USD: Format Code 001, and currency (Generic Unit) USD ZZZZ9: Format specifier. The input is a pattern, which is a combination of the Format Code and the Generic Unit (in this case this represents a Currency). In case the format specified cannot be found for this pattern, an error is logged. Now the request is to make it possible to log only an error in case the Format Code is not available at all. Workaround No workaround Test Procedure The following 3GL script, is used for testing: function main() { string result(80) | default currency of testcompany used is USD |the datecurr file contains the following entry: A|902|USD|001|001| |so, this defines default currency = USD for company 902. The default format code is 001. result = sprintf$("%A001", 1234.56) | no error, because the following entry exists in the datecurr file: C|001USD|USD|ZZZZ9| result = sprintf$("%A001,hfl", 1234.56) | Format Code 001 exists, but Generic Unit (currency) hfl does not exist | (the following line should be in datecurr: C|001hfl|hfl|ZZZZ9| ,but is not there) | in case hide_format_errors = 1, no error should be logged, because format code 001 itself exists | in case hide_format_errors = 0, error should be logged (old behaviour) result = sprintf$("%A999,HFL", 1234.56) | Format Code 999 does not exists | (the following line should be in datecurr: C|999USD|USD|ZZZZ9| ,but is not there) | in case hide_format_errors = 1, error should be logged (error specifies also the missing format code) | in case hide_format_errors = 0, error should be logged (old behaviour) } Resource hide_format_errors set to 0 ($BSE\lib\defaults\all) ============================================================ So, now the old behavior applies. When running the testscript the following errors are logged (this example shows the errors which would be logged on a Windows platform, but similair error messages will be logged in the ${BSE}/log/log.bshell file in case of Unix systems): Env Baan5c (C:\baan\baan5c) Prog bshell file \general\lib\al_1\AL_ENV.C # 933 Keyword get_format_entry Username kvddool type S language 2 Process 0xca4 ProcessSet kvddool@CNL09659:3196 Lasterror 0 Errno 0 bdberrno 213 (Transaction is started but not updated) Message get_format_entry pattern '001hfl' index 3 entry 0 not found Env Baan5c (C:\baan\baan5c) Prog bshell file \general\lib\al_1\AL_ENV.C # 933 Keyword get_format_entry Username kvddool type S language 2 Process 0xca4 ProcessSet kvddool@CNL09659:3196 Lasterror 997 (Overlapped I/O operation is in progress.) Errno 0 bdberrno 213 (Transaction is started but not updated) Message get_format_entry pattern '999HFL' index 3 entry 0 not found Resource hide_format_errors set to 1 ($BSE\lib\defaults\all) ============================================================ So, now the modified behavior applies, and only errors are logged in case the format code does not exists at all. When running the testscript the following errors are logged: Env Baan5c (C:\baan\baan5c) Prog bshell file \general\lib\al_1\AL_ENV.C # 941 Keyword get_format_entry Username kvddool type S language 2 Process 0xc94 ProcessSet kvddool@CNL09659:3248 Lasterror 0 Errno 0 bdberrno 213 (Transaction is started but not updated) Message get_format_entry pattern '999HFL' index 3 entry 0, format code 999 not found Like you see now, only an error is logged in case the Format Code at all cannot be found. The error message mentiones no also explicitly "format code xxx not found". Motive source DF:212983 MaintCorelli: # 17564 (BDUX11045): Adding functions to the shared library Date: Thu, 13 Mar 2003 11:26:28 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) The following functions were needed for external usage: 1. set_bse 2. get_bse 3. TSStssmbs_n2unicode_n 4. TSSunicode_n2tssmbs_n Test Procedure refer to mail fix text of BDNT11021 MaintCorelli: # 17563 (BDUX11048): Blat, partially backport of BDUX10786 to keep versions compatible Date: Thu, 13 Mar 2003 11:25:28 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) Binary Logging and Tracing: In porting set 7.3x and higher the relation between queries in the database drivers and the 3GL query has been lost. Since the introduction of the new query processor, the client-qid is not known anymore in the database drivers. All db queries get client qid -1. Test Procedure Code-changeble blat: sql s 01111 th off Run some sessions and verify the queries using bls6.2 and btop6.2 Type bugfix Motive source BRD:0-00000 MaintCorelli: # 17554 (BDUX9490): Wrong byte swap on Solaris Date: Thu, 13 Mar 2003 08:42:39 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Technical terms) A signed temporary variable was used iso an unsigned. Test Procedure See customer problem. Motive source TCS 100-131044 MaintCorelli: # 17536 (BDNT10963): Be able to handle install.info file in Unix style text format and removed full update/new components only dialog Date: Tue, 11 Mar 2003 12:25:35 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Installer comes with error popup if install.info file is in Unix style text format Problem Description (Technical terms) see above Workaround - Open install.info with wordpad (write.exe) - Save file in DOS style text format This workaround will only work if the medium is writable. Test Procedure Run installer, install update cd (be sure that the file update\install.info is in Unix style text format) and check that the installer do not popup with a message that the locale could not be retrieved from CD. MaintCorelli: # 17497 (BDUX10973): AS400: Repair CPlusPlusProgramTarget rule Date: Fri, 7 Mar 2003 08:16:02 +0100 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) The CPlusPlusProgramTarget rule had fault in $(CPPC) -c line. Test Procedure Nightly build MaintCorelli: # 17478 (BDUX10942): Tru64 oracle driver linked with version 8.1.7.4 Date: Thu, 6 Mar 2003 10:37:57 +0100 Created on: MaintCorelli Type: porting specific change Problem Description (Customer terms) Core dumps in Oracle driver on Tru64 when a lot of data is processed. Problem Description (Technical terms) Crash in Oracle 8.1.7.2 libraries. Motive source DF:133878 MaintCorelli: # 17475 (BDUX10919): do not ignore errors during release build Date: Thu, 6 Mar 2003 10:32:15 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Technical terms) wrong builds are delivered Workaround do not ignore error messages Test Procedure create an error in a sourcefile. try to build a portingset make should stop immediately after compilation error iso ignoring it and proceed with next sources. MaintCorelli: # 17414 (BDUX10880): AS400: Do bugf_BDUX10811 also for MaintCorelli Date: Fri, 28 Feb 2003 14:36:12 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) AS400 build faild because of: Yacc source /view/nightbuild.7.1.eddie.as400/vobs/tt/lib/qp/qc_qpc.y has been modified Please, update /view/nightbuild.7.1.eddie.as400/vobs/tt/lib/qp/qpc_yy_aix.c Yacc source /view/nightbuild.7.1.eddie.as400/vobs/tt/lib/qp/qc_qpc.y has been modified Please, update /view/nightbuild.7.1.eddie.as400/vobs/tt/lib/qp/qp_token_aix.h Workaround cp /vobs/obj.dbg.AIX43/lib/qp/qpc_yy.c /vobs/db/lib/qp/qpc_yy_aix.c cp /vobs/obj.dbg.AIX43/lib/qp/qp_token.h /vobs/db/lib/qp/qp_token_aix.h Test Procedure Weekly build MaintCorelli: # 17408 (BDUX10556): Sometimes no results returned with informix level 2 database driver. Date: Fri, 28 Feb 2003 10:31:24 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) sometimes no result are returned even it should with informix level 2 driver Problem Description (Technical terms) According informix documentation 'Handling a Parameterized SELECT Statement' the last byte of sqldata is reserved for \0 character. This character was not set by out driver. Unexpected results can occur during compare hostvariables if this character is not set. Workaround n/a Test Procedure This problem is difficult to reproduce, because it is depending what is stored in the memory area after sqlda.data. Hence you can reproduce it on system where the problem occured. MaintCorelli: # 17398 (BDUX10693): dividing by zero does not result in any bw-message on the ping Date: Thu, 27 Feb 2003 11:53:44 +0100 Created on: DevEscher Type: bugfix Problem Description (Technical terms) dividing by zero does not result in any bw-message on the ping, the result of the expression is 0, and the program continues. Test Procedure Use a script which dividing by zero on AIX bshell. Motive source TCS: 106-132431 MaintCorelli: # 17392 (BDUX10857): add include file Date: Wed, 26 Feb 2003 15:48:02 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) OS390 build failed after BDUX10629 fix MaintCorelli: # 17380 (BDUX10839): minor install bug Date: Tue, 25 Feb 2003 12:14:36 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) double check on exitance of bw.z typo in error message: intentet iso intended MaintCorelli: # 17378 (BDUX10815): History commands are overwritten in Debugger History Date: Tue, 25 Feb 2003 12:12:22 +0100 Created on: MaintReger Type: bugfix Motive source TCS:106-135190 MaintCorelli: # 17366 (BDUX10816): compilation error Date: Mon, 24 Feb 2003 11:33:07 +0100 Created on: MaintReger Type: bugfix Problem Description (Technical terms) compilation error Test Procedure compile MaintCorelli: # 17360 (BDUX10812): Delivery temporary dir fix Date: Mon, 24 Feb 2003 09:49:18 +0100 Created on: MaintReger Type: bugfix Problem Description (Technical terms) NT delivery leaves empty temporary directories. Workaround Delete them. Test Procedure Do a delivery on NT systems. MaintCorelli: # 17354 (HYNT10723): MSQL: Error in Hint generation Date: Mon, 24 Feb 2003 09:41:13 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Customer encountered dbs_error=1170, Microsoft SQL Server error 170, syntax incorrect. Problem Description (Technical terms) When the msql_opt_rows resource has value greater than 0 and ordered hint is given, the msql driver generates empty table hint. Eg. " with ()" Workaround set msql_opt_rows:0 in db_resource file. Alternatively remove msql_opt_rows from db_resource file. Motive source DF:133676 MaintCorelli: # 17344 (BDUX10776): new get.event.attribute.size function Date: Fri, 21 Feb 2003 13:15:22 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) get.event.attribute.size function required for new Webtop version Test Procedure Use new Webtop MaintCorelli: # 17339 (BDUX10513): Invalidation of db driver cache Date: Fri, 21 Feb 2003 09:24:44 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Baan error 512 (corrupt DD) on tables in a company that have been reconfigurated for another, similar, company. Problem Description (Technical terms) The database driver cache stores DD images per (physical) company. When a reconfiguration is done, only the DD's for the reconfigured companies have been invalidated in the cache. Other companies that with the same VRC are not invalidated. Workaround Refresh shared memory after reconfigurations. Test Procedure 1. Restart shared memory % shmmanger6.2 -s should show: Db driver DD 0x00000000 Note: run all following steps with DBSLOG=1000 and check the generated log. 2. Touch ttaad050 compnr 0 % qptool6.2 -c 0 -q "select * from ttaad050" dbs.log contains lines like: Search DBS_DD in shared memory for 'ttaad050000' - Not found Msg_type 8 received. Handle DD of 'ttaad050' compnr 000. Store dbs_dd in shared memory, key = 'ttaad050000' - hash created 3. Repeat step 2, and verify that the cache is used now. dbs.log contains lines like: Search DBS_DD in shared memory for 'ttaad050000' - Found 4. Repeat step 2, for another company number % qptool6.2 -c 1 -q "select * from ttaad050" dbs.log contains lines like: Search DBS_DD in shared memory for 'ttaad050001' - Not found Msg_type 8 received. Handle DD of 'ttaad050' compnr 001. Store dbs_dd in shared memory, key = 'ttaad050001' 5. Change the DD (remove a reference) and reconfigure table. % bdbreconfig6.2 -N ttaad050 -C 000 dbs.log contains lines like: Invalidate DBS_DD in shared memory. tname = 'ttaad050' - invalidated 6. Repeat step 2, verify that the cache is not used anymore. dbs.log contains lines like: Search DBS_DD in shared memory for 'ttaad050000' - Found but invalidated Note: run all following step with DBSLOG=1001 and check the generated log. 7. Repeat step 4, verify that the cache is not used anymore. dbs.log contains lines like: Search DBS_DD in shared memory for 'ttaad050001' - Found but invalidated An older database driver without this fix would pickup a DD from the cache and use it. The Dictionary dump shows that the old driver is still using an old DD. The dbs.log of the old driver contains lines like: Search DBS_DD in shared memory for 'ttaad050001' - Found MaintCorelli: # 17336 (BDUX10734): QP: comparison with 0.0 inconsistent between level 1 and 2 Date: Fri, 21 Feb 2003 09:22:19 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) In the WHERE clause of queries, the literals "0" and "0.0" behave differently. Problem Description (Technical terms) The double literal 0.0 is interpreted as a double with a precision of 1 digit after the decimal point. As a consequence, the query processor will evaluate the expression "0.0 = 0.04" to True. In constrast, the expression "0 = 0.04" yields False. Also, on level-2 comparing 0.0 to 0.04 yields False. The behavior is inconsistent between level-1 and level-2. Comparing 0.0 to 0.04 should yield False. Workaround Either change "0.0" to "0" or add more 0's after the decimal point. Test Procedure Execute the following command: qptool6.1 -q "select * from dbtst120 where 0.0 = 0.04" The query should return "No rows selected", because 0.0 does not equal 0.04. It should _not_ return all rows of table dbtst120. Motive source DF:131631 MaintCorelli: # 17317 (BDUX10655): Sequent oracle driver linked with version 8.1.7.4 Date: Thu, 20 Feb 2003 11:53:10 +0100 Created on: MaintCorelli Type: porting specific change Problem Description (Customer terms) Oracle 8.1.7.0 libraries return data when NULL values should be returned. Problem Description (Technical terms) Oracle bug solved in 8.1.7.1 Motive source DF:212621 MaintCorelli: # 17293 (BDNT10725): Do not start tgbrg4260m000 in case of Webtop login Date: Wed, 19 Feb 2003 12:46:35 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) When the DEM Browser is set as Startup program in the User Data, it is not possible to login using Webtop. Problem Description (Technical terms) The bshell detects whether the startup program equals tgbrg4260m000. If this is the case, it directly starts tgbrg4260m000. This way of working is introduced in the past with project VHUX2585, to solve a security leak. Later on, a extra check is built in to check if the user logs on using Worktop. Refer to project BDNT8097. Because Worktop has its own integrated DEM Browser, tgbrg4260m000 should not be started when logging in via Worktop. Now, it shows out also an exception has to be made when logging in via Webtop. Webtop does not support tgbrg4260m000, because it has also its own integrated DEM Browser. So the check is extended now specially for Webtop. (By checking for presence of (o)tmboaasmserver). Workaround Do not choose DEM Browser as startup program in your User Data. Webtop does "ignore" the setting of the startup program. Set "Worktop Tree" and "Baan Process Browser" to get the DEM process in Webtop. Test Procedure Set the startup program in your User Data to "DEM Browser", and convert to Runtime. Now try to log in using Webtop. You should have no problems now. Motive source DF:133693 MaintCorelli: # 17292 (BDUX10752): Backported project BDNT10428 to Corelli Date: Wed, 19 Feb 2003 12:39:53 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) Sockets were passed in strings in network byte order, where host byte order is expected, by e.g. James client (and developers). Numbers in host byte order converted to a string are much better readable/interpretable by users... See project BDNT10428 for more information. MaintCorelli: # 17287 (BDUX10747): NT installlers default to ISO_BIN for Single Byte installs Date: Wed, 19 Feb 2003 11:07:39 +0100 Created on: DevBaanIVc Type: porting specific change Problem Description (Technical terms) For single byte installs on NT ISO_BIN is a better locale than ISO-8859-1. Workaround Just do not choose the default :-) Test Procedure Start installing the full install CD MaintCorelli: # 17264 (BDUX10696): deliver error Date: Tue, 18 Feb 2003 10:13:23 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Technical terms) Error message during build when ~/.netrc does not exist Workaround Create a ~/.netrc file MaintCorelli: # 17257 (BDUX10713): merge error Date: Tue, 18 Feb 2003 09:01:23 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) Merge Error MaintCorelli: # 17239 (BDNT10632): USER variable ignored on Windows platform Date: Fri, 14 Feb 2003 16:18:02 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) We have a Windows account called "Test1". We have a Baan user called "test2". This Baan user has windows account "test1" as operating system user. So, in the utest2 file there is the following line: unixuser:test1 Now, when we log on to Baan, we have the following BW configuration: Username: test1 Password: Command: -- -set USER=test2 When we log in Baan, we are not logged on as "test2", but as user "test1". This is not correct. When we change the Windows username from "Test1" to "test1" this problem does not appear anymore, and you are logged in as test2. Because we are in a migration process from Unix to Windows we have capitals in the windows usernames. (Windows usernames are case-insensitive, but Unix usernames are case-sensitive) Problem Description (Technical terms) A case-insensitive compare is done between the LOGNAME (windows os user), and the username from the userfile. So in this case between "Test1" and "test1". The result is 0. Therefore a second, but case-sensitive string compare is done. In case both strings are not the same, the USER variable is set to test1. This should not be done by the bshell, instead do only a conversion to lowercase. (Because Baan Tools does not allow capitals in the Baan user name, nor in the Operating System user name). Workaround Take care that the Windows username is completely in lowercase. Test Procedure Refer to Problem description. Execute these steps, and see if you are logged in with the correct user. Motive source DF:133531 MaintCorelli: # 17237 (BDUX10670): Let Boot not decide to take a configuration if working on clearcase Date: Fri, 14 Feb 2003 16:16:37 +0100 Created on: DevEscher Type: porting specific change Problem Description (Technical terms) Enhanced Util/Boot script for ClearCase platforms. Do not ask for a choice Test Procedure Run Util/Boot in the object tree MaintCorelli: # 17228 (BDUX10677): Fixed compilation errors after changes for DEC 5.1a port Date: Fri, 14 Feb 2003 16:00:37 +0100 Created on: MaintBaanIVc Type: porting specific change Problem Description (Technical terms) Get HPUX10.20 compilable again with old java version MaintCorelli: # 17224 (BDUX8825): ftp portingsets Date: Fri, 14 Feb 2003 15:46:06 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) CAT does not accept rcp to their server, we have to use ftp. To be able to check the transfer, we also send a sum file. Test Procedure make deliver MaintCorelli: # 17220 (BDUX10505): Backspace emulation introduced for horizontal tab positions in filter6.X Date: Fri, 14 Feb 2003 15:39:05 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) Printout is not according to report layout on direct printer Unisys B9246-7. Problem Description (Technical terms) This printer is not able to print backspaces. We are using backspaces to set horizontal tab positions in lines. Workaround n/a Motive source DF:132816 MaintCorelli: # 17207 (BDUX10609): dividing by zero does not result in any bw-message Date: Thu, 13 Feb 2003 15:55:32 +0100 Created on: DevEscher Type: bugfix Problem Description (Technical terms) dividing by zero does not result in any bw-message on the ping, the result of the expression is 0, and the program continues. Test Procedure Compile a script which one is dividing by zero. Motive source TCS:106-132431 MaintCorelli: # 17200 (BDUX10658): Save diskspace. Do not store DOs for some targets Date: Thu, 13 Feb 2003 15:06:37 +0100 Created on: DevEscher Type: porting specific change Problem Description (Technical terms) Do not store too much data (derived objects) in clearcase. It's generated each build. MaintCorelli: # 17190 (BDNT10652): Changes in write_cd for Sun, -f added in read_tape_write_cd Date: Thu, 13 Feb 2003 11:24:52 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) Cpio on a SUN needs some more flags Test Procedure Run this script on a SUN machine (bimas for instance) MaintCorelli: # 17189 (BDNT10574): NT installer also needs locale choice for single byte Date: Thu, 13 Feb 2003 11:23:34 +0100 Created on: DevBaanIVc Type: bugfix Problem Description (Customer terms) Problem Description (Technical terms) After the changes of the NT installer, a locale choice is only asked for Multi Byte environments. For Single Byte the locale must also be chosen Test Procedure NT installation with new media MaintCorelli: # 17186 (BDUX10654): Fixed typo in rules file Date: Thu, 13 Feb 2003 09:49:29 +0100 Created on: DevEscher Type: bugfix Problem Description (Technical terms) Fixed typo. Use consistent names Test Procedure Build, and check that NormalSharedLibraryTarget3 (e.g. used for bclm) will pass the right options to COMPDATE tool MaintCorelli: # 17184 (BDUX10629): Ported DEC5.1a Date: Thu, 13 Feb 2003 09:47:56 +0100 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) Dec 5.1a port Workaround NA Test Procedure Build on host galaxy MaintCorelli: # 17171 (BDNT10287): Remove warning Language DLL not found Date: Wed, 12 Feb 2003 13:59:10 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) The Event Log of Windows contains a lot of Warnings like: Language DLL resBWj0.dll not found in directory C:\Program Files\Baan\fcbaan5c\bin\res, taking default (English) language Because of this Warnings real errors are more difficult to find. Problem Description (Technical terms) Warning issued in case resource DLL is not available. Remove warning. Workaround Refer to scopus solution 120944 Test Procedure Have for example a non-english BaanERP environment. Eg. Japanese. Log on to this environment. No warning like described in the Problem Description should be logged. Motive source DF:120944 MaintCorelli: # 17170 (BDUX10624): 3GL location in blat not correct when threshold logging is on Date: Wed, 12 Feb 2003 13:54:07 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) A BLAT trace sometimes shows "ottstp_stddll / __overloaded.exit" as the 3GL location of a query. This location is wrong. Problem Description (Technical terms) Under certain conditions the 3GL location was determined too late. Workaround Disable threshold logging, or add the 'sql 02' flag. Test Procedure Enable blat tracing 'sql s 011 th off' and run some sessions. The function __overloaded.exit should not be there anymore. MaintCorelli: # 17163 (BDNT9168): Conversion to Uppercase in Thai locale Date: Wed, 12 Feb 2003 13:50:14 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) The contents of the "To" field in Baan sessions shows an unknown character (a square), when the locale of the Baan user is set to ISO8859B or ISO_BINB. (Thai). Problem Description (Technical terms) The content of these "To" fields are filled by the bshell function fmax($#). This function determines the highest possible value according the current domain. In this case the domain is of the datatype string. When having the Thai locale, this results in character 251. So character 251 is the highest character for the Thai. Secondly a conversion to Uppercase is done on this character. And according the file \logic\mb\iso8859f.c the uppercase variant of character 251 is 219. But character 219 is a non-valid character for the Thai codepage (CP 874). Therefore it is necessary to overrule the standard uppercase conversion table. Workaround Do not do an uppercase conversion. Test Procedure 1. Have your windows configured for Thai (Default locale should be Thai) 2. Configure BW, tab NLS, also for Thai 3. Have your Baan User Locale set to ISO_BINB 4. Log in Baan and check the content of the "To" field, for example in session tcmcs0403m000 Print Warehouses (Baan5) It should contain a serie of characters 251. (Using tracing of BW: Show Events & Show Messages, Output to Console, you can see if character 251 is send). Motive source DF:130644 MaintCorelli: # 17160 (BDNT9923): Japanese error message not correctly displayed or logged Date: Wed, 12 Feb 2003 13:46:29 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Errormessages containing Japanese text are not displayed correctly in the BW Message Box. The Japanese part of the error message text is corrupted. The same problem applies for the error message, logged in the logfile (like log.bshell) or Windows Event Log. Also in this case the Japanese part of the text is corrupted. This is the case eg. for the error message "index out of dims" (MR_INDEXOUT). This error message is translated to Japanese, and this translated version is read from the file $BSE/lib/textinf/mrtext0j by the bshell. When this japanese translated error is displayed in the BW Message Box or logged it is no correct Japanese anymore. Problem Description (Technical terms) Regarding BW Message Box: Function ds_send_message replaces each % for %%. (Because % sign is a special sign for printf command). But in case of multibyte strings, one of the bytes of the complete multibyte string, can be such a % sign. In this case the % sign should not be replaced for %%, otherwise it will corrupt the whole multibyte string. Regarding logging: >From function mul_bsh_error, function log_printf() is called as part of the logging. This function log_printf() should do a conversion from TSS to native, before doing the actual logging. This is not done. Therefore TSS coded string is written to the logfile. In case of non-ascii characters it will result in carbage. (In most cases log_mesg() is used to write a log message. In that case the conversion from TSS to native is done.) Workaround Copy the english mrtext02 to mrtext0j. In this way you get english translated error messages instead of japanese translated. But they are at least readable. (if you can read the english language) Test Procedure To test you need to reproduce an error, which uses an error description from a japanese mrtext0j. Like the error MR_INDEXOUT = Index out of dims. I used the following 3GL scripts, which generates the error: Index out of dims: function main() { domain tcapi.type test2 long numar(20) test2 = ltoe(1) numar(0) = etol( test2 ) numar(11) = 10 message( "aaa" ) } 1. Log in a to a japanese BaanERP system, having the file $BSE/lib/textinf/mrtext0j 2. Run the testscript to generate the "Index out of dims" error 3. Check that the japanese translation of this error is correctly displayed in the BW Message Box, and also correctly logged in the $BSE/log/log.bshell file. (or in case of Windows: in the Windows NT Event Log). Motive source DF:131502 MaintCorelli: # 17158 (HYUX10636): Informix:No out of range error with date -1:TCS-100/130521 Date: Wed, 12 Feb 2003 13:39:57 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) The informix driver inserts the date 2 January 0001 when an attempt to insert the date -1 is made. All other drivers return error 205. Motive source TCS:100-130521 MaintCorelli: # 17145 (BDUX10563): missing oracle client file Date: Wed, 12 Feb 2003 10:41:24 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) $(ORA_HOME)/oracore/zoneinfo/timezone.dat is missing when oracle driver is build with Oracle9 Test Procedure install Baan MaintCorelli: # 17137 (BDUX10600): Informix level-2: Wrong bind of short values Date: Wed, 12 Feb 2003 09:33:19 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Technical terms) The Informix level-2 driver does not bind correct values to parameters of type short if the bind values are outside the range of short values. As a consequence, the Informix level-2 driver may return an incorrect result for a query, if the query contains an expression in which a column of type integer is compared to a constant or parameter which value is less than the minimal integer value (-32768) or greater than the maximum integer value (32767) For example: select empno from dbtst120 where edlevel inrange -100000 and 100000 Column edlevel is of type integer. The Informix level-2 driver returns no rows for this query. It should return all rows of table dbtst120, because -100000 is less than the minimal integer value and 100000 is greater than the maximum integer value. Workaround In the query, use the minimum and maximum integer values. Test Procedure Create a file 'query' that contains the following query: select empno from dbtst120 where edlevel inrange -100000 and 100000 Use qptool to execute this query: qptool6.2 -c 0 -f query Execute this query on a level-2 Informix driver. Verify that it returns all rows of table dbtst120. Motive source DF:133420 MaintCorelli: # 17127 (HYNT10614): MSQL: Disable join hints by default Date: Tue, 11 Feb 2003 16:11:53 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) By default join hints are enabled. Enabling join hints can give huge performance problem (as observed at the beta customer site) Workaround Disable the join hints by setting the resource msql_join_hint:NONE or the environment variable MSQL_JOINT_HINT:NONE Test Procedure No join hint should be seen in the query printed in dbs.log MaintCorelli: # 17123 (BDUX10286): Query Timeout Implementation in backend job Date: Tue, 11 Feb 2003 10:42:05 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) Manual merge of project HYUX10361. Since corelli supports OS400 Version V4R5 also and cancelling of query execution is not possible in V4R5, version check is done for the implementation. Test Procedure Run a query through BSQL with -t option and check whether query execution is stopped, in case of timeout. MaintCorelli: # 17114 (BDUX10114): Fix BREAK button in debugger Date: Tue, 11 Feb 2003 08:48:55 +0100 Created on: DevEscher Type: bugfix Problem Description (Technical terms) For session fmlbd0110m000 I have the dal fmlbd010 in debug. When session is started I want to start debugging after the next action, so Break is pressed. However the debugger never breaks/stops. Select PLN002795, Break, and then press toolbar button 'actualize plan' Workaround Set debug flag on in the process list. Test Procedure See the Problem Description Motive source TCS:106-130690 MaintCorelli: # 17071 (BDUX10390): Removed dead code (SQLserver 6.5 import) from installer and portset Date: Thu, 6 Feb 2003 09:43:19 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) After the media replacement the alternative way to import SQLServer dumps is not used anymore. Test Procedure Full install from the new media MaintCorelli: # 17068 (BDNT10532): Buffer too short in text handling Date: Thu, 6 Feb 2003 09:06:06 +0100 Created on: DevBaanIVc Type: bugfix Problem Description (Customer terms) Japanese BaanIV environment crashes on full-length label (264 characters) Problem Description (Technical terms) The $BSE/lib/TEXTINFO files are read per line or 255 bytes when the line is longer. Because the japanese language update contains a description of the maximum allowed length (264 bytes) only part of the description is read. The remaining part was cosidered as a new but illegal entry, causing a bshell termination. Workaround truncate lines in $BSE/lib/TEXTINFO/* files to 255 bytes. Also truncate descriptions in table ttadv060, as the files can be converted to runtime from these files. Test Procedure Do a full installation of BaanIV with japanese language update. Change language of the user from '2' (english) to 'j' (japanese) Start BW. The original bshell terminates, the new one works fine. MaintCorelli: # 17059 (BDUX10512): the debugger evaluates expression when pressing right mouse button on variable. Date: Wed, 5 Feb 2003 14:29:08 +0100 Created on: MaintReger Type: bugfix Problem Description (Technical terms) the debugger evaluates expression when pressing right mouse button on variable. Test Procedure when pressing a variable in the debugger with the right mouse button the variable is shown in the top window of the debugger, this is easy for changing the variable a little before evaluating (adding "/u" for utc-fields or "(2)" for array-variable). no evaluation on right mouse button click Motive source TCS Call: 106-134358 MaintCorelli: # 17054 (BDUX10520): Fixed compile error on DEC 5.1 Date: Wed, 5 Feb 2003 13:42:21 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) Fixed compile error for DEC 5.1 MaintCorelli: # 17046 (BDUX10389): DB2: drop table is not possible if there are cursors open on the table in Level1 Date: Wed, 5 Feb 2003 09:17:03 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Drop table gives a FLOCKED error running in LEVEL1 mode. Problem Description (Technical terms) DB2 does not allow dropping a table with open cursors on it. Test Procedure Run the RDD test case under CTSID104 and TCID100. Motive source TCS:100-127620 MaintCorelli: # 17045 (HYUX10493): Informix:Error -1210 with pdaemon6.2:TCS-100/128821 Date: Wed, 5 Feb 2003 09:16:41 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) Starting pdaemon6.2 fails with INFORMIX error -1210. Test Procedure 1. Start pdaemon6.2 2. Run the following gcommand script #1 ttaad320 TCHORD 2 #1 ttaad320 REQUAL 0 -i -ENOREC = ttaad320.stat = 12 = ttaad320.date = 0 With a driver that does not contain this fix, both test cases will fail with error 2210. MaintCorelli: # 17040 (BDUX10511): Change logof of bshell to include old server Date: Tue, 4 Feb 2003 17:13:17 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) If you have a network licence environment it can get a condition the licences are not released. Problem Description (Technical terms) if the license deamon that give the license to the bshell is no longer the first in the list licence6.x file the license is not released when the bshell is shutting down. Workaround shutdown of the license deamon that has the license available, or change the secuence of servers in the list. Test Procedure use a license in a network license server. start a bshell te retrieve a license, then change the sequence number of the license file so the other server is the first to check. then logout of to release the license again. check with licmon if the license is released. Motive source DF:133372 MaintCorelli: # 17028 (BDUX10502): Installer bug in BDNT10037: wrong ExistsDir test Date: Tue, 4 Feb 2003 11:59:34 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Installer does not show dialogue for choosing update CD Problem Description (Technical terms) Wrong test on directory existence. MaintCorelli: # 17020 (BDUX10487): DF211808 Date: Mon, 3 Feb 2003 17:17:26 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) When bshell goes down (abnormally or) with SESSION_TIMEOUT reached then all subprocesses (of the bshell) except apix.x goes away. Additionally, apix.x uses message queues which are still active when it's parent processes are gone. Problem Description (Technical terms) apix.x is started from the application via run.baan.prog function. When processes are started this way, the bshell opens up a sh_server to manage the external application. When the bshell goes away, it takes care of it's child processes, but not it's grandchildren. Likewise, sh_server does not currently have a mechanism to remove it's child processes. Additionally, apix.x will still have active message queues even if it is killed. So these resources are not freed up. A later fix may be realized in the run.[baan.]prog function to dictate the behavior of child (of sh_server) processes in case of abnormal exits. Workaround 1) (if possible) do not kill bshell, but exit normally 2) do not use SESSION_TIMEOUT to kill processes which use apix.x 3) manually kill apix.x process; and remove message queues Test Procedure Start an application session that will start apix.x. Leave session open so that the apix.x remains active. Tested by "killing" the bshell. Also tested by allowing bshell to timeout vie SESSION_TIMEOUT set. Verify process apix.x goes away. Also verify that message queues are gone. Motive source DF:211808 MaintCorelli: # 17010 (BDUX10462): DB2: Limiting the scope of the fix BDUX9864 to the import session Date: Mon, 3 Feb 2003 15:43:35 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) The fix BDUX9864 was made specific to AIX platform. This fix limits the scope of the fix further to the import session. Test Procedure Run the import session on AIX/DB2 and see that it runs faster. Motive source TCS:100-128182 MaintCorelli: # 17004 (BDUX10464): Changed internal separator to be able to handle files with spaces Date: Mon, 3 Feb 2003 09:11:53 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Tape -> cd project could not be completed, because script failed on filenames with spaces Problem Description (Technical terms) Internal file separator was a space, causing not able to run constructions like for i in `find . -print` because space acted as separator. Test Procedure create a cd from tape, be sure that a file with spaces is existent MaintCorelli: # 16996 (BDUX10471): Modification in comments Date: Fri, 31 Jan 2003 14:32:35 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) Fixed compilation problem caused by a typo Test Procedure Recompiled and verified the fix MaintCorelli: # 16995 (BDUX10365): AS400: Port MaintCorelli to OS version V5R1 Date: Fri, 31 Jan 2003 13:18:53 +0100 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) OS400 version V5R4 is no longer supported by IBM, so the minimal OS400 version is now V5R1. Workaround Use -qTGTRLS=V5R1M0 as compiler flag and GENCSRC instead of #pragma mapinc Test Procedure Build MaintCorelli on a V5R1 AS/400 with the correct ibm_as400V5R1.cf file. MaintCorelli: # 16993 (BDUX10339): Missing bind values in trace log Date: Fri, 31 Jan 2003 13:02:14 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) If TT_SQL_TRACE option is specified with value 02000, generally the trace log contains the value(s) of the bind parameter(s) specified in the query, along with other information. However, when the value of the bind variable remains unchanged for more than one bind, the trace log contains values only for the first bind. For other bind(s), the value(s) of the bind variable(s) is missing from the trace log. Test Procedure QPTool was tested with the following options: TT_SQL_TRACE=02000 qptool6.2 -c 0 -apefbefbc -f "query file". The query file contained the query: "select user, name from ttaad200 where user = :user". This means that QPTool was run with actions intending to have two fetches for the value specified in the bind variable "user".: The following are snippets of the trace log, after the fix. Tracing on: TT_SQL_TRACE = 02000 [SqlParse: before parse] SqlId : 40066008 - QID : 1 - Status : 0 Trans 1 (qptool) Mode 0 Query: select user, name from ttaad200 where user = :user [SqlParse: after parse (e = 0)] SqlId : 40066008 - QID : 1 - Status : 2 [SqlExec before exec: (CompNr 000)] SqlId : 40066008 - QID : 1 - Status : 2 Automatic Bind for user Value(str): 'gerardv' ... [SqlExec before exec: (CompNr 000)] SqlId : 40066008 - QID : 1 - Status : 7 Automatic Bind for user Value(str): 'gerardv' [This line was missing for second or more binds] ... MaintCorelli: # 16992 (HYUX10463): Informix:Fixing date handling problems in DEC Date: Fri, 31 Jan 2003 13:01:24 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) Date handling problems Test Procedure None MaintCorelli: # 16981 (BDUX10443): create file after dir creation Date: Thu, 30 Jan 2003 17:38:22 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Technical terms) command file is created in non existing dir. Workaround mkdir /vobs/general/porting/tmp Test Procedure Relnotes MaintCorelli: # 16954 (BDNT10404): Trace menu-data structure Date: Thu, 30 Jan 2003 11:39:20 +0100 Created on: DevEscher Type: bugfix Problem Description (Customer terms) Tracing DsNmenuData structures sent by the backend is not done in BW. This makes analyzing problems in this area very difficult. Problem Description (Technical terms) DsNmenuData structure is shown as one block of binary data or not at all when it is too large. A function must be added which traces these attributes in a more readable format. Test Procedure Switch-on message tracing in BW and check the log file. MaintCorelli: # 16909 (BDUX10376): Fixed some minor issues in object tree Date: Tue, 28 Jan 2003 12:05:15 +0100 Created on: DevCorelli Type: porting specific change Problem Description (Technical terms) Cleanup object tree after enabling clearcase on meson (DEC OSF 4.0f) Workaround ignore the error MaintCorelli: # 16906 (HYNT10200): L1 to L2 Migration Tool fails when a table has > 11 hash columns Date: Tue, 28 Jan 2003 11:14:13 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) L1 to L2 migration tool fails when a table has more than 11 hash columns Workaround Drop the hash columns manually for the failed tables and reexecute the migration tool. Test Procedure Execute the tool on a table with more than 11 hash columns and it should succeed. Motive source TCS:100-129723 MaintCorelli: # 16905 (BDUX10371): Adjusted text in installation script for DB2 Date: Tue, 28 Jan 2003 09:58:07 +0100 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) Adjusted text in script (remark found during Linux test) Test Procedure run db2v5_install6.2 and check the first screen. The reference to the log file for errors should be correct now. MaintCorelli: # 16877 (BDUX10112): Bshell did not support advanced OLE-DB connector calls from OpenWorld Date: Fri, 24 Jan 2003 16:00:41 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) OLE-DB Connector of OpenWorld was not able to handle table aliases, or to handle arrays. Problem Description (Technical terms) Metadata information, provided by Bshell, did not contain enough information, so the OLE-DB connector was not able to map this to it's internal datastructures. Motive source DF:289483 MaintCorelli: # 16867 (BDUX10181): Fixed Problems in BDBLibraries Date: Thu, 23 Jan 2003 19:16:35 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) This fix comprises of backporting changes of the following project: DevEscher: # 15854 (BDUX9294): Fixed problems in BDB Libraries. For more details please refer the mail fix text of BDUX9294. MaintCorelli: # 16851 (HYNT10305): Database List in Installer Screen Date: Thu, 23 Jan 2003 10:50:20 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Installer Screen shows Informix version 9, but document says Informix version 9 is not supported Workaround Choose Informix version 9, to install BaanERP in Infomrix 7.31. Test Procedure Run the installer with new medium and check whether screen shows Informix Dynamic Server. Motive source TCS:100-129268 MaintCorelli: # 16844 (BDNT10297): Disabled alpha on Windows in imagebuilder Date: Wed, 22 Jan 2003 16:18:17 +0100 Created on: MaintBaanIVc Type: porting specific change Problem Description (Technical terms) Disabled Alpha support in image builder MaintCorelli: # 16842 (BDNT10296): Fixed compilation bug in imagebuilder Date: Wed, 22 Jan 2003 16:16:46 +0100 Created on: MaintBaanIVc Type: porting specific change Problem Description (Technical terms) Wrong prototype in source Test Procedure compile source file, and check that no warnings/errors will occurr. MaintCorelli: # 16803 (BDNT10146): cd.read6.x scripts uses LANG variable Date: Mon, 20 Jan 2003 15:06:00 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) cd.read6.x script uses LANG variable, SunOS platform gives error message on using this LANG variable as an internal variable. Changed the variable in LANGUAGE. Test Procedure Use cd.read6.x on a SunOS system, if you do not see the error message: "couldn't set locale correctly" the scripts works ok MaintCorelli: # 16784 (BDUX9990): Performance optimization: only make shlib with PIC Date: Mon, 20 Jan 2003 10:29:00 +0100 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) Adding compiler flags for PIC (Position Independant Code) caused a performance decrease. As PIC objects are only needed for shared libraries, and not for the majority of the porting set (most tools are linked statically), it has been decided to compile the libraries in PIC mode only when required. In those cases a non-PIC variant is made too. So we end up with 2 technically different but functionally identical libraries when these are needed to create a shared library. Test Procedure Nightly build MaintCorelli: # 16779 (BDNT10213): Changed text in dialog Date: Mon, 20 Jan 2003 09:06:52 +0100 Created on: MaintCorelli Type: porting specific change Problem Description (Customer terms) Adjusted text in installer dialog, to prevent confusion Problem Description (Technical terms) Adjusted text in installer dialog, to prevent confusion MaintCorelli: # 16773 (BDUX9864): DB2:altering tables to be VOLATILE for import Date: Fri, 17 Jan 2003 16:27:18 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Import runs very slowly on DB2 on AIX Problem Description (Technical terms) During import tables grow and the queries on these tables tend to become slower and slower as the import proceeds. Because the statistics for these tables are also not updated DB2 seems to take a wrong query plan. Workaround Run update statistics on some problem tables when the import runs. Test Procedure Test with this driver and see that the import runs faster on AIX. Motive source TCS 100-128182 MaintCorelli: # 16737 (BDUX10170): Fixed nightly build due to BDNT9825 Date: Thu, 16 Jan 2003 08:51:22 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) Nightly build failed after merge of BDNT9825 due to BaanIV BaanERP differences Test Procedure Just build MaintCorelli: # 16726 (BDUX10028): Followup on BDUX9325/9526 handle corrupt bdbreconfig dump Date: Wed, 15 Jan 2003 12:10:09 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) This change replaces the projects BDUX9325 and BDUX9526 Running bdbreconfig6.1 -N+ on corrupt dump, moves the dump. Before project BDUX9325 and BDUX9526 the dump was deleted. When a reconfigure (bdbreconfig) is executed on a table, and something goes wrong, a file R.
must saved on the system. This dump can be imported using the command: bdbreconfig6.1 -N
+ However, when this command is executed on a "corrupt" dump, a appropriate error must be givven, and the data must be preserved to be able to do some kind of manual restore. Problem Description (Technical terms) The change in project BDUX9325 and BDUX9526 to move the dump was on second though not a right choice, as a second problem would destroy the data as wel, and a manual intervention is needed anyway. Test Procedure Create a file called R.tccom000 containing just some characters. Then start bdbreconfig -Ntccom000+. This should fail with an appropriate message as the dump file is corrupt. The corrupt dump file should not be removed by bdbreconfig. Motive source DF:130255 MaintCorelli: # 16722 (BDNT10149): TSS: bad merge of project 9859 Date: Wed, 15 Jan 2003 09:49:32 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) bad merge of project 9859 Motive source DF:129448 MaintCorelli: # 16709 (BDNT9859): TSS: add mb.locale.enumerate Date: Tue, 14 Jan 2003 17:36:44 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Technical terms) info about locales needed, especially about their character sets Motive source DF:129448 MaintCorelli: # 16695 (BDNT10068): cd.read contains a construction, which is not valid for sun Date: Tue, 14 Jan 2003 16:01:33 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) cd.read contains a construction, which is not valid for sun MaintCorelli: # 16687 (BDUX9967): date.num() - 1 on Informix Date: Tue, 14 Jan 2003 13:45:23 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) Baan SQL queries with an expression on a date field will return a wrong results on an Informix database driver in Level 2 mode. Problem Description (Technical terms) The next query (see Test Procedure) will return a wrong results on an Informix database driver in Level 2 mode. The query does not use the correct field type when it sends the query to Informix database. Workaround Example where clause: tdsls045.invd >= date.num() - 1 Workarround would be: tdsls045.invd - date.num() >= -1 Test Procedure Execute the next query and check if it returns the correct results: select orno, pono, srnb, invd from tdsls045 where tdsls045.invd >= date.num() - 1 Motive source DF:131351 MaintCorelli: # 16684 (BDNT9825): Licensing service does not automatically start during restart of system Date: Tue, 14 Jan 2003 13:32:44 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) When the Windows system is restarted, the Licensing service does not start automatically. This problem is specially occurring on Windows 2000 o.s. The Licensing services uses functionality of the windows operating system to retreive the MAC addresses of the networkcards. The MAC address is one of the unique identifiers, used for creating the brandinfo. During the restart of the Windows operating system, it appeared that the MAC addresses could not be retreived, because certain parts (netbios interface) of the Windows o.s. is not fully initialized yet. Because the working of the Licensing service is depending on this, the Licensing service did not start correctly. Problem Description (Technical terms) The netbios function, used in mid() does not return (keeps hanging). This happens only when Licensing service is started again during restart of the system. Probably the underlying OS part (NetBios interface) is not initialized at that moment, and causing this behaviour. Workaround Start the Baan Licensing service Manually. Motive source DF:131473 MaintCorelli: # 16671 (BDNT10120): Running of a batch (.bat) file has a problem Date: Tue, 14 Jan 2003 08:54:48 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) After customer upgraded the Grieg portingset (6.2a.x.x.) a batchfile does not work correctly anymore. The contents of the batchfile is, for example: @echo off rem Set paramaters needed for each run set BSE=C:\Progra~1\Baan\fcbaan5c set INPUTD=%BSE%\tmp FOR /F "TOKENS=1* DELIMS=" %%A IN ('TYPE %INPUTD%\test_input.txt') DO SET LOGDATE=%%A echo ----->>%INPUTD%\test_output.txt echo %LOGDATE%>>%INPUTD%\test_output.txt The contents of the file test_output.txt is: Parameter_TEST When running this batch file from a dos window, a file test_output.txt is created. The content of this file will be: Parameter_TEST But, when running this batchfile from Baan (via session ttaad5205m000 Add Shell Command to Job), the %LOGDATE% seems to be not filled. After running the batchfile in this way, the content of test_output.txt is: ECHO is off. Problem Description (Technical terms) Some creation flags for CreateProcess in sh_server were added by project BDNT8015. (Delivered with portingset 7.1c.05). When removing flags DETACHED_PROCESS (no console), and CREATE_NO_WINDOW (no console window), the batchfile is working correctly again. Workaround No workaround Test Procedure 1. Have a batchfile as described in the Problem Text 2. Have also a file test_input.txt as described in Problem Text 3. Run this batchfile via a job, by adding this Shell Command to a Job (ttaad5205m000) 4. When executed once, a file test_output.txt should be created containing: Parameter_TEST Motive source DF:132671 MaintCorelli: # 16640 (BDUX10085): Nightly build failed Date: Fri, 10 Jan 2003 15:20:40 +0100 Created on: MaintBaanIVc Type: porting specific change Problem Description (Technical terms) Util/Boot failed because type does not exist in sh during cronjob. MaintCorelli: # 16637 (BDNT10037): Do not ask for installation source when installing BW Date: Fri, 10 Jan 2003 15:16:08 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) When installing BW, using setup.exe, a window pops up asking you for the Installation Source. (master cd or update cd). This window is not applicable for a BW installation, and therefore should not appear. Problem Description (Technical terms) Some modifications in the setup.rul were needed, to get correct flow in case of BW install. Workaround Select the default option (master cd) and continue. Test Procedure Run installation of BW. For each step do also try if Back button works. MaintCorelli: # 16614 (BDNT9952): Merged changed for deliver.exe (BDNT9903) to Corelli Date: Thu, 9 Jan 2003 11:06:06 +0100 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) Deliver did not: - build the cameo/ddc examples in the right way - some settings from Reger were mistakely added in deliver - the content.txt (bill of materials) file was not filled in any case (bug in windows?) Workaround - use an older deliver version, e.g. 2.27 - disable automatic build of ddc/cameo examples in deliver.cfg Test Procedure Run deliver with building cameo/ddc on. Check the content.txt file, and check that e.g. cameo1.exe was created in the destination path (under $DEST\bshell\api\bin) MaintCorelli: # 16612 (HYUX10029): Not all linked cursors closed when a table is dropped:TCS-100/127635 Date: Thu, 9 Jan 2003 10:23:19 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) All cursors linked to a table are not closed when a table is dropped. Because of this an attempt to access a table soon after it is dropped and re-created causes IDS to return error -710 (Table has been dropped, altered or renamed on an insert). Test Procedure Run the following gcommand script. #1 ddcts542 TCRTBL 0 #1 ddcts546 RINS 0 -i -E REF_NOT_EXISTS = ddcts546.int = 3 = ddcts546.int1 = 3 = ddcts546.int2 = 0 #1 ddcts546 DABORT 0 #1 ddcts545 TDRTBL 1024 #1 ddcts545 TCRTBL 0 #1 ddcts546 RINS 0 -i = ddcts546.int = 1 = ddcts546.int1 = 1 = ddcts546.int2 = 0 #1 ddcts546 DCOM 0 A driver that does not contain this fix, will return error -710 with the second RINS. Motive source TCS:100-127635 MaintCorelli: # 16589 (BDNT10008): Manual merge e-boks updates to Corelli maintenance branch Date: Tue, 7 Jan 2003 15:41:06 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) To get e-boks available for Corelli on short term, the projects: BDNT9402 and BDNT9608 must be merged to the Corelli maintenance branch as well. Problem Description (Technical terms) Initially these projects where only merged to the Corelli development branch. However due to the delivery schedule of BW, to get this functionality out as soon as possible, it must also be included in the maintenance branch. Test Procedure Test E-box on Corelli. MaintCorelli: # 16570 (BDUX9888): DB: Error 201 on delete after update Date: Tue, 7 Jan 2003 11:54:52 +0100 Created on: DevCorelli Type: bugfix Problem Description (Customer terms) An error 201 occurs when splitting the complete quantity of a shipment line from shipment A to shipment B. Problem Description (Technical terms) A delete or update after an update within the same transaction can result in error 201 (E_ROW_CHANGED), if the (first) update updates a double value. Test Procedure The following gcommand script can be used to test. It first insert a double value of 1.4. It then updates this value to something close to 0.0. Then it deletes the row. dbtst120 RINS 0 -i = dbtst120.empno = 21 = dbtst120.salary = 1.4 = dbtst120.workdept = A00 * * Put a delayed lock on salary 1.4 * dbtst120 REQUAL# 0 -i = dbtst120.empno = 21 * * Update the salary from 1.4 to 2.2204460492503131e-16 * The value in the Delayed Lock Buffer (DLB) is updated to * this new value. * dbtst120 RUPD% 0 -i = dbtst120.salary = 2.2204460492503131e-16 * * Delete the row. The DLB value (2.2204460492503131e-16) was * compared with a number close to it, resulting in error 201. * dbtst120 RDEL% 0 * * Verify that the row is deleted (results in flush of the delete). * dbtst120 REQUAL 0 -i -E NOREC = dbtst120.empno = 21 Motive source DF:129680 MaintCorelli: # 16566 (BDNT9977): Adjusted text in installer dialog for locales Date: Tue, 7 Jan 2003 10:43:36 +0100 Created on: MaintCorelli Type: porting specific change Problem Description (Customer terms) Adjusted text in locale dialog in installer. Made a reference to the documentation. Without reference, it might be unclear for customers to choose the right locale. Problem Description (Technical terms) see above. Test Procedure run installer and check text in dialog for choosing the locale. MaintCorelli: # 16561 (BDUX9838): Assign values to Long goes wrong Date: Tue, 7 Jan 2003 10:38:04 +0100 Created on: MaintReger Type: bugfix Problem Description (Technical terms) BIC compiler doesn't accepted the -2147483648 value for long variables as Programmers Guide mentioned. Test Procedure Compile a 3GL source with long = -2147483648 constant. Motive source TCS:106-128610 MaintCorelli: # 16500 (BDUX9931): adjust copyright string Date: Thu, 2 Jan 2003 11:39:12 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Technical terms) Copyrightstring is changed from 2002 to 2003 MaintCorelli: # 16497 (BDUX9842): If number of logical companies > 100 refint gives a n error message Date: Thu, 2 Jan 2003 10:05:22 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) When running refint6.x an error message appears saying 'Syntax error: 'selec' not expected'. Refint6.x keeps on running, but the error message also keeps on comming once every number of times. Problem Description (Technical terms) The problem lies in the fact that the number of logical companies linked to a physical company for specific tables exceeds the defined maximum, which is 100. This maximum number of 'parent_compnrs' is defined within the refint sourcefile refint.c. When the number of logical comanies exceeds this maxmimum a wrong memory location is read, with unpredictable results. Workaround There is not realy a good workaround. Only thing what can be done is decreasing the number if logical->physical links in the compnr6.x file. Test Procedure Create a compnr6.x file with more that 100 logical->physical definitions to a table which is checked directly or indirectly by use of refint6.x. On the Corelli environment on Melmac this can be tested using the following command: refint6.2 -r -Ntccom001 -C903. Motive source DF:131987 MaintCorelli: # 16490 (BDUX9920): incorrect Makefile Date: Mon, 30 Dec 2002 16:16:55 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) Define causes problems with make, clearmake works well. Workaround Use clearmake Test Procedure make deliver MaintCorelli: # 16486 (BDUX9918): Permission check on company switch not correct for reports Date: Mon, 30 Dec 2002 15:17:19 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) The next simple line is working correct for a supper user, however this is not working for a normal user if this is part of a report script. it is working fine when this is executed in a normal program script: ret = db.bind("ttcmcs002","",902) Problem Description (Technical terms) Permissions of the normal user are checked based on the session that is used. if the user is allowed to start this session in the new company then the actions is done correct. this is however giving a problem if this is part of a report script, because on reports no authorizations are set. so this can never be found. Workaround do not use this construction in a report, or let it be executed a supper user. Test Procedure create a report with the next part in the report script: ret = compnr.check(902) |or other company number ret = db.bind("ttcmcs002","",902) |or other table/companynumber take a normal user with all possible permissions and execute this part of the code. check the return values and see if the return value is correct. Motive source DF:211385 MaintCorelli: # 16473 (BDUX9891): incorrect classpath when setting BAANHOME Date: Tue, 24 Dec 2002 08:34:35 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) The path for finding java classes (com_baan_bclm.jar) was set to ${BAANHOME} instead of ${BAANHOME}/java/com_baan_bclm.jar when the BAANHOME envronment variable was set. Problem Description (Technical terms) The path for finding java classes (com_baan_bclm.jar) was set to ${BAANHOME} instead of ${BAANHOME}/java/com_baan_bclm.jar when the BAANHOME envronment variable was set. Workaround Set the correct path for finding the java classes in the bse_vars or .bwc file. Test Procedure set BAANHOME Start tmboaserver session. MaintCorelli: # 16466 (BDNT9883): cd.read6.x: stopped prematurely while creating install.info, selecting of tools package not handled correctly Date: Mon, 23 Dec 2002 13:43:11 +0100 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) cd.read6.x: stopped prematurely while creating install.info when only installing applications selecting of tools package not handled correctly when the tools package had the same vrc another package Workaround Allways start installing with a portingset cd, always select everything Test Procedure Do start installing a applications cd from that cd see if the script ends with: cd.read6.1: Ended If so, it's ok MaintCorelli: # 16441 (BDUX9828): AS/400: Convert binary files to code page 819 before using pax Date: Fri, 20 Dec 2002 13:49:11 +0100 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) Since building the portingset using the native build method instead of the distributed build method, binary objects in directory $BSE/lib are incorrect. This is due to the fact that the code set of the files was unequal to 819 before they were archived in the tar file. Because they are untarred in code set 819 some character conversion would occur. The first native build was used in 7.1C.04. Workaround Convert all files to code page 819 before putting them in the tar file. Test Procedure Build a portingset on AS/400, pax -v -C 819 -r -f OS4004.5_DT.PROTO.tar, pax -v -C 819 -r -f EXTRA.TAR, ftp lib/unicode/SHIFTJIS.N2U in binary mode to an unix machine and diff it against the unix version. MaintCorelli: # 16434 (BDUX9735): Global compile drops the resulting objects in a LOWER vrc Date: Fri, 20 Dec 2002 08:18:28 +0100 Created on: DevEscher Type: bugfix Problem Description (Technical terms) Global compile drops the resulting objects in a LOWER vrc if it doesn't have the permissions to create them in the correct VRC Workaround Improve the al_1 lib's sopen() function to be capable recognize a writing on a non writeable object, and don't let it to do more explode (fallback). Test Procedure give an object non-write permissions in 7.4a compile it in 7.4a see the result not appear in 7.3_a_sp1 Motive source TCS:106-128592 MaintCorelli: # 16416 (BDUX9844): bad fix in BDUX9467 Date: Thu, 19 Dec 2002 11:59:37 +0100 Created on: DevEscher Type: bugfix Problem Description (Customer terms) Stack traces are displayed incorrectly sometimes after BDUX9467 Found by Jan Koen Annot Problem Description (Technical terms) sometimes the variable names are displayed incorrectly. Test Procedure 3gl session in debug and display a stack trace MaintCorelli: # 16400 (BDNT9805): modify identing with quotes Date: Tue, 17 Dec 2002 16:54:53 +0100 Created on: DevEscher Type: bugfix Problem Description (Technical terms) The text after #ident needs quotes around it. Workaround where identstring is put, put quotes too (in deliver.exe) Test Procedure running modified deliver.exe with modified deliver.lst MaintCorelli: # 16391 (BDUX9467): Bshell crash during mul_bsh_error stack trace dump Date: Tue, 17 Dec 2002 10:10:19 +0100 Created on: DevEscher Type: bugfix Problem Description (Customer terms) Bshell crashes from time to time during the logging of an error Problem Description (Technical terms) the bshell stackframe is incomplete during some of the mul_bsh_error calls During the dumping an illegal symbol is accessed causing the bshell to crash. MaintCorelli: # 16369 (BDNT9790): al_amem.h: do not include macros.h Date: Mon, 16 Dec 2002 09:13:11 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) compile problem on Sequent Test Procedure compile on Sequent MaintCorelli: # 16353 (BDNT9776): Fix BOM file in NT deliver Date: Fri, 13 Dec 2002 08:56:06 +0100 Created on: DevEscher Type: bugfix Problem Description (Customer terms) BOM file were empty when COMPRESS-ed arhcive were created in the NT delivery. Problem Description (Technical terms) The destination path weren't found in the destination path of the copied file, because it were copied into a temporary place before compression. Workaround When compression flag is set, check for the temporary dir path also. Test Procedure Do a deliver, and check the BOM file's content. MaintCorelli: # 16347 (BDNT9771): Make all variants of deliver on all releases identical Date: Fri, 13 Dec 2002 08:52:29 +0100 Created on: MaintReger Type: bugfix Problem Description (Customer terms) Deliver were not identical on every releases Problem Description (Technical terms) Because of an enhancement project, a change weren't merge to all releases. Workaround Do it manually. Test Procedure Build a deliver executable. MaintCorelli: # 16309 (BDUX9721): javapath corrupt Date: Wed, 11 Dec 2002 16:52:02 +0100 Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) Upgrade the portingset on an HP system and you can no longer login or run tmboaserver using bshelljava6.1. Bshell6.1 works fine. Logging shows the CLASSPATH is not set properly. Problem Description (Technical terms) HPUX JRE1.1.8.00 and JRE1.1.8.05 JDK1_1InitArgs structure have different sizes. Compiletime structure has a smaller size than the runtime version. Memory is corrupt when JNI_GetDefaultJavaVMInitArgs has been called. Workaround Install JRE1.1.8.05 and use HPUX10.20 bshelljava. Test Procedure tmboaserver Motive source DF:211110 MaintCorelli: # 16290 (BDUX9730): AS/400: building BSHTMPLIB.a results in UPDBNDDIRE error Date: Wed, 11 Dec 2002 09:34:28 +0100 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) Because sha and base64 routines were added to the bshell the command line for (q)ar command became too long which resulted in a UPDBNDDIRE error while building BSHTMPLIB.a. Workaround On AS/400, create BSHTMPLIB.a in two steps. Test Procedure Nightly build MaintCorelli: # 16283 (BDNT9697): Fix some problems on NT delivery Date: Tue, 10 Dec 2002 13:56:19 +0100 Created on: DevEscher Type: bugfix Problem Description (Customer terms) Deliver.exe doesn't give error messeage, when the example compilations failed or a variable were not defined. Problem Description (Technical terms) msdev.exe doesn't give an error code back when it fails to open a project file. Workaround Put the error messages to standart error and check for compilation success. Test Procedure Do a nightly build with DDC and CAMEO examples enabled. MaintCorelli: # 16265 (BDNT9699): XML: improve NULL pointer protection Date: Fri, 6 Dec 2002 16:11:58 +0100 Created on: MaintReger Type: bugfix Problem Description (Technical terms) bshell crashes when 3GL passes value 0 to xmlFindNodes Workaround test arguments in 3GL Test Procedure see TCS remark Motive source TCS:106-130838 MaintCorelli: # 16250 (BDNT9552): version.txt handling added in Unix to cd scripts Date: Fri, 6 Dec 2002 10:08:15 +0100 Created on: MaintCorelli Type: bugfix Problem Description (Technical terms) version.txt handling added in Unix to cd scripts Workaround fill version.txt by hand MaintCorelli: # 16238 (BDUX9687): partly merge of port_VHUX4305 Date: Thu, 5 Dec 2002 16:17:29 +0100 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) Compilation error on PTX4.4.6 MaintCorelli: # 16212 (BDUX9664): BIC crash fix Date: Wed, 4 Dec 2002 13:50:53 +0100 Created on: DevEscher Type: bugfix Problem Description (Customer terms) Bic crashes after compiling this: function boem( long k, f ) { } function boem( long k, long l ) { } Problem Description (Technical terms) The error recovering doesn't cleans up after an error has been found. Workaround Don't let the BIC work with the corrupted (function) data after recovering. Test Procedure Compile the given source example. Motive source TCS:106-128119 MaintCorelli: # 16203 (BDUX9322): BIC question mark problem fix Date: Tue, 3 Dec 2002 15:12:26 +0100 Created on: MaintReger Type: bugfix Problem Description (Customer terms) In tisfcdll0226 I had a construction like: (condition ? value1: value2) As this gave a compilation error ("value1 not expected") Problem Description (Technical terms) BIC recognizes value1: as a label and not a variable. Workaround Check labeling inside ? condition, and don't allow that. Test Procedure Compile the given example. Motive source TCS:106-128733 MaintCorelli: # 16140 (BDUX9607): Increment BW version for 7.1d.02 Date: Thu, 28 Nov 2002 13:37:05 +0100 Created on: MaintCorelli Type: porting specific change Problem Description (Technical terms) New version number due to delivery of 7.1d.01 Test Procedure check bshell6.2 -v Total number of changIDs displayed: 160