List of changes for portingset: 6.1c.07.11 MaintBaanIVc: # 30382 (BDNT22293): Make automation more robuust to prevent hangup with Worktop Date: Fri, 14 Apr 2006 10:27:07 +0200 (CEST) Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) Internal problem. Test Procedure Use Worktop to communicate with BaanIV backend. Affected Executables bw.exe MaintBaanIVc: # 30305 (BDUX22328): function creat_tmp_files() can fail Date: Fri, 7 Apr 2006 17:34:12 +0200 (CEST) Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) In some situations function creat_tmp_files$() can not create more then 1000 files. Further the performance is bad. Workaround n/a Test Procedure Run following 3gl test program function main() { long i for i = 1 to 1001 if creat.tmp.file$( "/tmp/jr") = "" then message("creat.tmp.files$() failed") endif endfor } Affected Executables bshell6.x Motive source SDIS:TECH-00000-00390 MaintBaanIVc: # 30188 (BDUX22228): resolve merge problem BDUX22193 Date: Mon, 3 Apr 2006 11:54:25 +0200 (CEST) Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Because of merge problems a declaration and initialisation of a variable was not done. Therefor the build process broke down. Workaround N/A Test Procedure Build the stuff Affected Executables Bshell MaintBaanIVc: # 30141 (BDUX22193): qss search function return value (repaired badfix) Date: Thu, 30 Mar 2006 15:53:59 +0200 (CEST) Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Previous fix of project BDUX21383 was badfix. It resulted in the fact that labels where not displayed in some sessions anymore. reason was a wrong return value after the qss.qsearch function in qss_fun.c. Workaround Use an older portingset Test Procedure Compile the following script in debug and check return values: function extern main() { long search.def(3,4) long a(6,3) long i |* Found index long br(1,3) |* To be found value. |* Fill search def qss.start( search.def, 1,1 ) qss.start( search.def, 2,2 ) qss.start( search.def, 3,3 ) qss.type( search.def, 1, DB.LONG ) qss.type( search.def, 2, DB.LONG ) qss.type( search.def, 3, DB.LONG ) |* Fill the array with values a(1,1) = 12 a(1,2) = 2 a(1,3) = 2 a(2,1) = 12 a(2,2) = 20 a(2,3) = 20 a(3,1) = 12 a(3,2) = 20 a(3,3) = 20 a(4,1) = 12 a(4,2) = 20 a(4,3) = 20 a(5,1) = 12 a(5,2) = 20 a(5,3) = 20 a(6,1) = 12 a(6,2) = 200 a(6,3) = 200 br(1,1) = 12 br(1,2) = 1 br(1,3) = 1 i = qss.search( QSS.EQLE+QSS.SRC.IS.SORTED+QSS.SRC.DUPL.ALLOWED, br, |* value to find a, |* Array search.def) br(1,1) = 12 br(1,2) = 10 br(1,3) = 10 i = qss.search( QSS.EQLE+QSS.SRC.IS.SORTED+QSS.SRC.DUPL.ALLOWED, br, |* value to find a, |* Array search.def) br(1,1) = 12 br(1,2) = 20 br(1,3) = 20 i = qss.search( QSS.EQLE+QSS.SRC.IS.SORTED+QSS.SRC.DUPL.ALLOWED, br, |* value to find a, |* Array search.def) br(1,1) = 12 br(1,2) = 20 br(1,3) = 10 i = qss.search( QSS.EQLE+QSS.SRC.IS.SORTED+QSS.SRC.DUPL.ALLOWED, br, |* value to find a, |* Array search.def) br(1,1) = 12 br(1,2) = 199 br(1,3) = 199 i = qss.search( QSS.EQLE+QSS.SRC.IS.SORTED+QSS.SRC.DUPL.ALLOWED, br, |* value to find a, |* Array search.def) br(1,1) = 12 br(1,2) = 999 br(1,3) = 999 i = qss.search( QSS.EQLE+QSS.SRC.IS.SORTED+QSS.SRC.DUPL.ALLOWED, br, |* value to find a, |* Array search.def) } Affected Executables bshell MaintBaanIVc: # 30131 (BDNT22080): Collate flag is not read from printer setting by BwPrint Date: Thu, 30 Mar 2006 15:18:27 +0200 (CEST) Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) When the collate printer setting for a printer is default set off and a report is sent to the windows printer from the Baan application, BwPrint does not get the collate setting from the printer. Default the collate function in BwPrint is enabled (when the chosen printer supports collating). Workaround Set the collate flag manually. Test Procedure 1. First try to setup your local windows printer (select any windows netwrok printer). In the properties of this windows printer, Go to printing preferences, Advanced, and select 2 copies of paper and UNCHECK Collate button, and press ok and apply to save. 2. In Baan try to have the same device added. 3. Go to any print session, press print, then you have ttstpsplopen, here set 2 copies and enable the preview option. Once you continue, you get a preview screen, Now press the print button. Check if the option COLLATE is UNCHECKED. When there is no collate- checkbox in the print dialog look at the properties of the printer. Affected Executables BwPrint.exe Motive source SDIS:TECH-00000-00619 MaintBaanIVc: # 30096 (BDUX22031): PAM support for Unix Date: Wed, 29 Mar 2006 10:02:45 +0200 (CEST) Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) PAM only available on HPUX-PARISC Affected Executables bshell6.1 badmin6.1 Motive source DF:227579 MaintBaanIVc: # 30018 (BDNT21905): BW error message window cannot be maximized Date: Fri, 24 Mar 2006 10:02:48 +0100 (CET) Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) The BW Error Message Window can be resized, but the maximize option (icon top right corner of the window and in the pull down menu) is disabled. Workaround N/A Test Procedure Start any environment using BW. Make sure an error will occurr e.g. wrong password. When the error message window appears try to maximize it using the mximize icon on the top right corner of the window. Affected Executables bw.exe Motive source SDIS:TECH-00000-00526 MaintBaanIVc: # 29922 (BDUX21964): Install: oracle 10g Date: Fri, 17 Mar 2006 15:34:15 +0100 (CET) Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) Windows install on Oracle 10g does not find ORACLE_HOME in the environment or registry Workaround Two solutions: 1) run setup.exe with the -nocheck option 2) Set the %ORACLE_HOME% environment variable for the setup.exe process. Test Procedure Windows install on clean (minimal installed) Oracle 10g. Affected Executables setup.ins setup.dbg MaintBaanIVc: # 29905 (BDUX21938): Install: oracle 10g not always recognized Date: Thu, 16 Mar 2006 10:56:05 +0100 (CET) Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Error during install on Oracle 10g: version not recognized Workaround Do not exit the installer, but hit "Resume" and continue the install. Test Procedure Install on Oracle 10g Affected Executables Windows installation MaintBaanIVc: # 29884 (BDUX21908): Fix project BDUX21383 undone Date: Tue, 14 Mar 2006 16:08:35 +0100 (CET) Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Labels not displayed after fix BDUX21383. only on none-dynamic form with multiple tabs. Test Procedure Run session tfgld1101m000 and look if the field labels are there. Affected Executables bshell6.x MaintBaanIVc: # 29868 (BDUX21902): qptool always dumps dlock info on stderr Date: Tue, 14 Mar 2006 10:13:22 +0100 (CET) Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) qptool6.1 -q "select . . . for update" generates a lot of unneeded trace information on stderr. Test Procedure Run any FOR UPDATE query with qptool and compare the output on stderr. Affected Executables qptool MaintBaanIVc: # 29824 (BDUX21820): Bshell: improved performance of delayed locking mechanism Date: Fri, 10 Mar 2006 12:23:59 +0100 (CET) Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) When a large number of rows is locked, using the delayed locking mechanism, performance could become less than optimal. Workaround Change the application logic to do additional commit.transaction() calls after a number of db.update()/db.delete() calls. This is only possible if the application logic allows to split the transaction. Test Procedure See the original changes Affected Executables bshell (all changes) all database drivers (BDUX9888) Motive source SDIS:TECH-00000-00344 MaintBaanIVc: # 29678 (BDUX21531): Startup sessions not started when bshell is started by script Date: Wed, 1 Mar 2006 12:22:08 +0100 (CET) Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) When startup sessions have been defined for a user these sessions do not start automatically when the bshell is started using a script. Test Procedure 1. Create a script for starting the bshell containing e.g. file bshell.test: #!/bin/ksh exec /trash/basram/bshell6.1 $@ 2. Make am entry in the $BSE/lib/ipc_info for redirection to the script e.g.: bshell_test s 0 0 p $BSE/tmp/bshell.test 3. Create a file in directory $BSE/lib/user e.g. test.sessions. In this file the sessions which should be started are mentioned. 4. Add the following to the userfile: bshell_test.sessions: test.sessions When the bshell is started the sessions mentioned in the file test.sessions should start automatically at bshell startup. Affected Executables ipc_boot bshell Motive source SDIS:TECH-00000-00264 MaintBaanIVc: # 29479 (BDUX21547): QP: TT_SQL_TRACE=20 corrupts result Date: Mon, 13 Feb 2006 12:09:55 +0100 (CET) Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) Setting TT_SQL_TRACE=020 (possibly in combination with other flags, for example 07777) results in corrupt query results. Workaround Don't set the 020 flag, or set RDS_FULL=1. Test Procedure Run a query with RDS_FULL larger than 1, and such that in the result set values repeating values. Run once with TT_SQL_TRACE=020 and once without and observe that the returned result sets are identical. Affected Executables qptool bshell all database drivers Motive source SDIS:TECH-00000-00408 MaintBaanIVc: # 29381 (BDUX21394): Never allocate level-1 row cache for level-2 drivers Date: Fri, 3 Feb 2006 16:38:30 +0100 (CET) Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) Level-2 drivers might allocate too much memory. Workaround For level-2 drivers explicitely set the field refresh_time to 0. Test Procedure regression tests Affected Executables all database drivers MaintBaanIVc: # 29341 (BDUX21383): return value position qss search incorrect Date: Wed, 1 Feb 2006 13:04:00 +0100 (CET) Created on: MaintCorelli Type: bugfix Problem Description (Customer terms) When an item is searched in an array using qss functionality the returned position is not correct when the item to be searched is not present in the array and the searchflags QSS.LESS or QSS.EQLE is used and duplicates are present in the array. Workaround Create an indexed array. Test Procedure Create the following 4GL script, compile and run it: |****************************************************************************** |* gatst9004 0 VRC B60 a |* |* sdvries |* 05-12-16 [12:40] |****************************************************************************** |* Script Type: 0 |****************************************************************************** #ident "@(#)gatst9004 gaB60a sprinter Rev.No. 2 16 Dec 05 sdvries" function extern main() { long search.def(1,4) #define MAX_NO 12 long a(MAX_NO) |* Fill search def qss.start( search.def, 1,1 ) qss.way( search.def, 1, QSS.UP ) |* Search upwards? qss.type( search.def, 1, DB.LONG ) |* Fill the array with values a(1) = 732292 a(2) = 733292 a(3) = 733292 a(4) = 733292 a(5) = 733293 a(6) = 733293 a(7) = 733293 a(8) = 733293 a(9) = 733295 a(10) = 733295 a(11) = 733295 a(12) = 733295 long i |* Found index long v |* To be found value. boolean found v = 733294 |* This value is not in the array.... i = qss.search( QSS.LESS+QSS.SRC.IS.SORTED+QSS.SRC.DUPL.ALLOWED, v, |* value to find a, |* Array search.def, MAX_NO) |* expect 733293 at position 8, 733293 is the first value |* that is less then 733294 and position 8 is the first |* position to be found if searched upwards. if i = 8 then |* last value in the array. found = true else |* No, no, this is not correct, the last value |* should be found. found = false endif |* Try it for an other value this is in the array. v = 733293 i = qss.search( QSS.LESS+QSS.SRC.IS.SORTED+QSS.SRC.DUPL.ALLOWED, v, |* value to find a, |* Array search.def, MAX_NO) |* expect 733292 at position 4, smaller then 733292. |* And the first value found if searched upwards (ascending). if i = 4 then |* last value in the array. found = true else |* No, no, this is not correct, the value 4 |* should be found. found = false endif } Affected Executables bshell Motive source SDIS:ERP-00000-00555 MaintBaanIVc: # 29323 (BDUX21396): Informix CSDK 2.90 for DEC platform Date: Tue, 31 Jan 2006 13:01:01 +0100 (CET) Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) Update to Informix CSDK 2.90 for DEC platform necessary. MaintBaanIVc: # 29289 (BDUX21170): Oracle 10g install for Windows Date: Thu, 26 Jan 2006 15:10:28 +0100 (CET) Created on: MaintCorelli Type: porting specific change Problem Description (Customer terms) Windows install media for BaanIV and Corelli will be created again. These new media must support Oracle 10g as well. Test Procedure Full install Affected Executables full windows install MaintBaanIVc: # 29284 (BDUX21339): UPDATE STATISTICS for Informix implemented Date: Thu, 26 Jan 2006 10:19:23 +0100 (CET) Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) To ensure that the Informix optimizer selects a query plan that best reflects the current state of your tables UPDATE STATISTICS is necessary. Workaround manual task of database administrator Affected Executables inf_srv6.1 MaintBaanIVc: # 29169 (BDUX21132): Password aging check (-checkpwd) failed on some AIX (5.3?) versions Date: Tue, 17 Jan 2006 17:13:14 +0100 (CET) Created on: MaintBaanIVc Type: porting specific change Problem Description (Customer terms) badmin6.x -chkpwd failed on getting age of password. Workaround Set all attributes, which is not really a workaround. We only found this problem on a customer (AIX5.3) system. Test Procedure Run badmin6.x -chkpwd and check that valid info (age etc) will be shown Affected Executables badmin6.x MaintBaanIVc: # 29122 (BDUX20339): DB: error 509 with mirrored database Date: Fri, 13 Jan 2006 11:37:39 +0100 (CET) Created on: MaintBaanIVc Type: bugfix Problem Description (Customer terms) Error 509 is issued on database actions if database mirroring is used. The chance on this error is bigger of some tables are mirrored and some other tables are not mirrored. Test Procedure See the regression tests (mirrored_cursors/001a & mirrored_cursors/001b) for reproduction scenarios. Affected Executables bshell Motive source TECH-00000-00227