List of changes for portingset: 6.1c.07.15



MaintBaanIVc: # 35362 (BDNT27146): Manual merge of no_ipc_msg_window

Date: Wed, 9 May 2007 10:34:53 +0200 (CEST)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
Running a lot of users on NT might result in not able to start 
sort.exe or sh_server.exe processes due to some Windows
limitation. Special registry tuning is needed to over come this.
See solution 121503 for more information.
This solution might still not acceptable for all customers because
the registry settings might conflict with requirements of other
applicatons.

Workaround
See solution 121503

Test Procedure
Take a windows system, set the third argument of SharedSection
(See the solution) to a low value (e.g. 256 of less). Disable 
combo-mode in tabledef and try to a number of bshells (e.g. the
baan benchmark) One will notice that 30 users will run fine,
but running 40 users will result in errors on the console due
to not able to start sort.exe.
Set the resource 'no_ipc_msg_window:1' in $BSE/lib/defaults/all
and notice that one can at least run upto 400 users on the same
system.

Affected Executables
ntbshell.exe, *_srv.exe, bshcmd.exe

Motive source
SDIS:TECH-00000-02172




MaintBaanIVc: # 35310 (BDUX27481): Authorization check error with non-existent 
OS user

Date: Fri, 4 May 2007 15:28:42 +0200 (CEST)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
User can logon even if unixuser in maintain user data session was set to a wrong or non-existent system user.

Motive source
SDIS:TECH-00000-02206




MaintBaanIVc: # 35244 (BDUX27412): permission check fails

Date: Thu, 26 Apr 2007 16:14:17 +0200 (CEST)

Created on:  DevChopin
Type:        bugfix

Problem Description (Customer terms)
I have made changes to the form of session tfgld1431m000.

I have next problem:
1. Start session tfgld1431m000 from Sessions (ttadv2500m000).
2. Close session tfgld1431m000.
3. Select tfgld1431m000 in Sessions and press Specific.
4. Select Edit Form.
5. Getting next error:

Fatal Error : [MR_NO_PERMS] No permission for ttstpoledaem

6. I have to kill Sessions to continue.


Test Procedure
See problem text

Affected Executables
bshell




MaintBaanIVc: # 34924 (BDNT27053): Make SQLserver resources conform 
documentation and enable EnvVars also via resource

Date: Fri, 30 Mar 2007 17:09:15 +0200 (CEST)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
See changes from branch bugf_BDNT27002

Workaround
See changes from branch bugf_BDNT27002

Test Procedure
See changes from branch bugf_BDNT27002

Affected Executables
msql7_srv.exe




MaintBaanIVc: # 34611 (BDNT26656): fill java.library.path property

Date: Thu, 8 Mar 2007 10:08:33 +0100 (CET)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
Starting tmboaserver via BW works fine.
1) Windows:
But when trying to start openworld via bshell in the background (started from command line or via ASM), it fails. The bshell is hanging.
Dbgjvmi shows the following lines:
B:0000026:::(00001):Exception occured in start (I)V
B:0000027:::(00001):Exception cleared

2) Linux:
When trying to start tmboaserver in the background (via commandline, or via ASM), it is needed to extra add $BSE/shlib to the LD_LIBRARY_PATH.
Otherwise tmboaserver doesn't start.

Result of this is, that the ASM server has to be hard linked to a specific 
BSE environment, and cannot be used anymore for multiple BSE environments.

* when ASM is started with:
	LD_LIBRARY_PATH=/usr/java/j2sdk1.4.2_06/jre/lib/i386:/usr/java/j2sdk1.4.2_06/jre/lib/i386/client or with unset LD_LIBRARY_PATH
error "Cannot start, failed to start JAVA, wrong JDK version"

* when ASM is started with:
	LD_LIBRARY_PATH=/usr1/baan4/bse/shlib:/usr/java/j2sdk1.4.2_06/jre/lib/i386:/usr/java/j2sdk1.4.2_06/jre/lib/i386/client
works fine

Workaround
In $BSE\java\jvm_options:
-Djava.library.path=<bse_dir>\shlib

On Windows:
append PATH with <jre_dir>\bin;<jre_dir>\bin\client in system environment variabeles
append PATH with %BSE%\bin;%BSE\shlib in environment variabeles of each BSE by means of baanman snapin
the ASM service will use these PATHs

On Linux:
append LD_LIBRARY_PATH with <bse_dir>/shlib in each $BSE/lib/bse_vars
append LD_LIBRARY_PATH with <jre_dir>/lib/i386:<jre_dir>/lib/i386/client:<jre_dir>/lib/i386/native_threads before starting ASM server from command line
before starting a bshell, ASM server will read its bse_vars file

Test Procedure
See Problem description

Motive source
TCS:800-156464
TCS:800-156465




MaintBaanIVc: # 34602 (BDNT26703): Add option to pass install directory to setup

Date: Wed, 7 Mar 2007 13:53:33 +0100 (CET)

Created on:  DevChopin
Type:        bugfix

Problem Description (Customer terms)
When using the unattended install for the BW client you are unable to select the directory
to install the new client in.
When running the BW setup.exe /s /u from the command line.
This defaults to install in Program files\baan\Baan ERP
There is no way to install it in another directory.

Workaround
Use attented install.

Test Procedure
See problem description.

Affected Executables
setup.exe

Motive source
SDIS:TECH-00000-02018




MaintBaanIVc: # 34584 (BDUX26565): password expired if baan login daemon used

Date: Tue, 6 Mar 2007 15:10:46 +0100 (CET)

Created on:  MaintCage
Type:        bugfix

Problem Description (Customer terms)
Baanlogin returns password expired even when it is not expired.

Workaround
Decrease Maximum Password Age.

Affected Executables
blogind6.x
badmin6.x

Motive source
SDIS:TECH-00000-02050




MaintBaanIVc: # 34473 (BDUX26570): Loop in call graph profiler

Date: Wed, 28 Feb 2007 13:25:16 +0100 (CET)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
Call Graph Profiler hangs during a session close

Test Procedure
start tdsls4102m000 and close
bshell shouldn't hang any more

Affected Executables
bshell6.1

Motive source
TCS:800-156455




MaintBaanIVc: # 34457 (BDUX26502): Enable Enhanched PAM error codes to be passed 
to the user.

Date: Tue, 27 Feb 2007 15:01:52 +0100 (CET)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
It is not always clear what kind of error is given by PAM when
changing passwords

Affected Executables
bshell6.1
badmin6.1

Motive source
SDIS:TECH-00000-01742




MaintBaanIVc: # 34314 (BDNT26393): BW Setup for Windows Vista

Date: Wed, 14 Feb 2007 16:07:46 +0100 (CET)

Created on:  DevChopin
Type:        bugfix

Problem Description (Customer terms)
BW Setup on Windows Vista fails.

Test Procedure
Try to install BW on Windows Vista.

Affected Executables
setup.exe




MaintBaanIVc: # 34297 (BDUX26383): Barcode texts need to escaped if quotes are 
in barcode text

Date: Wed, 14 Feb 2007 10:28:24 +0100 (CET)

Created on:  DevChopin
Type:        bugfix

Problem Description (Customer terms)
Barcode on unix not correctly printed if barcode contains double quote or dollar sign (or any other character interpreted by shell)

Workaround
Postpone barocode processing to $BSE/lib/barocde/type%d script. Do not pass non-ascii data in barcode (bc$()).

Test Procedure
Write a 3GL script which call bc$() with double quote in barcode caption.

Affected Executables
filter6.x

Motive source
SDIS:TECH-11111-99999 (no defect yet created)




MaintBaanIVc: # 34252 (BDNT26304): baan IV

Date: Fri, 9 Feb 2007 17:18:52 +0100 (CET)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
Incorrect optimizer directive generated by the driver for Level-1  queries.

Baan query

   select * from ticst001
   where _index3 > {""}
   order by _index3

will be translated to following database query

SQL> select {+ INDEX_ONE(ticst0015513abaan) } t_pdno, t_pono, t_cprj, t_sitm, t_
revi, t_opol, t_cwar, t_leng, t_widt, t_noun, t_qune, t_scpf, t_ques, t_cpes, t_
qucs, t_cpcs, t_cpcp, t_issu, t_subd, t_obqn, t_opno, t_aldt, t_bfls, t_nnts, t_
pnpt, t_Refcntd, t_Refcntu, hash3 from "baan".tticst001551 where hash3 > ? order
by hash3 ;

Index hint INDEX_ONE(<indexname>) is not a valid directive. It should be INDEX(<tablename> <indexname>)
instead.



Workaround
n/a

Test Procedure
Run above qptool query and check with DBSLOG=01770 the database query. Ensure inf_init:01524 is set in db_resource.


Affected Executables
inf_srv6.x




MaintBaanIVc: # 34149 (BDNT26236): Use infenv_lock_timeout for group sessions 
also

Date: Mon, 5 Feb 2007 09:15:13 +0100 (CET)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
During bdbreconfig sometimes error 113 (table locked) appears.

Workaround
n/a

Test Procedure
This is difficult to test. The problem was seen on big tables, but it was more a random
issue.

Affected Executables
inf_srv




MaintBaanIVc: # 34100 (BDNT26184): Informix level 2 error 215

Date: Wed, 31 Jan 2007 17:18:10 +0100 (CET)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
Informix level 2 customers are faced somtimes to 
error -215 (illegal state)

Workaround
Set inf_open_cursors:100 in db_resource file. If the problem still
exist increase the value until error 215 disapears.


Affected Executables
inf_srv

Motive source
SDIS:ERP-00000-14526




MaintBaanIVc: # 34059 (BDNT26145): MSQL: invalid cursors after deadlock recovery

Date: Tue, 30 Jan 2007 13:49:09 +0100 (CET)

Created on:  MaintCage
Type:        porting specific change

Problem Description (Customer terms)
On SQL2000: E_BDB_ABORT (error 850) with invalid MSQL cursors after returning to
a db.retry.point() on deadlock recovery.

Workaround
Upgrade to SQL2005

Test Procedure
Using 2 bshells run the following two scripts on SQL2000:

| BSHELL 1
table tdbtst100
table tdbtst120

function deadlock1()
{
	db.retry.point()

	if db.retry.hit() > 0	then
		message("User1: Jumped to retry point %d, e = %d",
			db.retry.hit(), e )
	endif

	select * from dbtst100 for update
	where deptno = "A00"
	selectdo
		dbtst100.deptname="HET ROMMELHOK"
		db.update(tdbtst100,db.retry)

		db.flush()
		message("User1: After 1st db.flush point  ")
	endselect

	select * from dbtst120 for update
	where empno = 10
	selectdo
		dbtst120.lastname = "KONIJN"
		db.update(tdbtst120,db.retry)

		db.flush()
		message("User1: After 2nd db.flush point  ")
	endselect

	commit.transaction()
}

function main()
{
	deadlock1()
}

| BSHELL 2
table tdbtst100
table tdbtst120

function deadlock2()
{
	db.retry.point()

	if db.retry.hit() > 0	then
		message("User2: Jumped to retry point %d, e = %d",
				db.retry.hit(), e )
	endif
	
	select * from dbtst120 for update
	where empno = 10
	selectdo
		dbtst120.lastname = "RABBIT"
		db.update(tdbtst120,db.retry)

		db.flush()
		message("User2: After 1st db.flush point  ")
	endselect

	select * from dbtst100 for update
	where deptno = "A00"
	selectdo
		dbtst100.deptname="DUMP STORE"
		db.update(tdbtst100,db.retry)

		db.flush()
		message("User2: After 2nd db.flush point  ")
	endselect

	commit.transaction()
}

function main()
{
	deadlock2()
}

- Start both bshells and the different sessions till both show the message "after 1st ..."
- continue the 1st bshell (it blocks)
- continue the 2nd bshell (it blocks too)
- a deadlock detection by the database will cause both to continue to the 2nd message.
- both will jump to the retry point. In the old situation one with e  = 850 and one with e = 201
- continue with the session having e = 201 and run till completion
- continue with the other (e = 850). This one keeps on failing and the event log shows invalid cursor errors

With this fix both sessions jump to the retry point (with e = 201) and do not get invalid cursors anymore.

Affected Executables
msql7_srv

Motive source
SDIS:TECH-00000-01859




MaintBaanIVc: # 33973 (BDNT26048): Wrong user permissions on BSE/tmp let crash 
bshell

Date: Tue, 23 Jan 2007 12:16:47 +0100 (CET)

Created on:  DevChopin
Type:        porting specific change

Problem Description (Customer terms)
Bshell crashes when trying to create a file in $BSE/tmp but failed due to strict permissions on file level.

Workaround
Set right permissions on $BSE/tmp (and $BSE/log for unix).

Affected Executables
bshell and others who use HAL_access_ex().