List of changes for portingset: 7.1d.11


MaintCorelli: # 28227 (BDNT20436): Solve problem in creat_tmp_file.

Date: Thu, 3 Nov 2005 15:44:22 +0100 (CET)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
Function creat.tmp.file$( [string pathname] ) fails in case:
the pathname contains ${BSE}, ${BSE_TMP} or a remote hostname (host!path).

This problem is detected in portingset 6.1c.07.09.
(before release, during testing)

Workaround
Do not use ${BSE} or ${BSE_TMP} or remote host syntax.

Test Procedure
Run the following testscript (3Gl):

function main()
{
        string  myfile(100)

        myfile = creat.tmp.file$()

        myfile = creat.tmp.file$("c:\baan\fcbaan4c\tmp")
        myfile = creat.tmp.file$("${BSE}\tmp")
        myfile = creat.tmp.file$("srv-nld-tls1!c:\temp")
        myfile = creat.tmp.file$("srv-nld-tls1!${BSE}\tmp")

	|wrong situatons
	|subdir. tem does not exits
        myfile = creat.tmp.file$("c:\baan\fcbaan4c\tem")
        myfile = creat.tmp.file$("${BSE}\tem")
	|system srv-nld-tas does not exits (although specified in remote-user file)
        myfile = creat.tmp.file$("srv-nld-tas!c:\temp")
	|subdr. tem does not exist on remote system
        myfile = creat.tmp.file$("srv-nld-tls1!${BSE}\tem")

	|wrong situatons
	|no write permision for subdir. privat
        myfile = creat.tmp.file$("c:\baan\fcbaan4c\privat")
        myfile = creat.tmp.file$("${BSE}\privat")

}

In case of the correct situations, myfile should contain the name of the tmp file.
In case of the wrong situations, myfile is empty, and an error is logged in the Event Log / bshell.log.

Note: The script above is created for testing on a Windows environment. For testing on a Unix environment,
you need to modify it.


Affected Executables
bshell, fs, etc.




MaintCorelli: # 28202 (BDUX20207): MySQL driver: Multi byte support, multi 
language support, binary_columns

Date: Wed, 2 Nov 2005 14:41:29 +0100 (CET)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
1. Multi byte support is not sufficient. When using the GB2312 locale the tables
   are created in the latin1 character set.

2. Special Chicago Metallics requirement: it is not possible to create the tables
   in utf8 and have one user operating in ISO88591 and another user operating in 
   GB2312.

3. MySQL has a default timeout on the connection of 28800 seconds (8 hours). This
   means that after 8 hours of inactivity the connection is automatically closed.
   It should be possible to specify the timeout via a resource.


Test Procedure
Specify the resources as specified and observe that the tables are created
in the correct character set (you may use DBSLOG=400 for that).


Affected Executables
mysql_srv6.2




MaintCorelli: # 28195 (BDUX20427): taudit dumps core on remote BSE setup

Date: Wed, 2 Nov 2005 11:18:19 +0100 (CET)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
taudit dumps core on $BSE_REM setup

Test Procedure
See remark

Affected Executables
taudit

Motive source
TCS:800-150050




MaintCorelli: # 28180 (BDUX20399): Windows badmin: add support for retrieving 
password expiration warning info (merge from BDNT18449)

Date: Mon, 31 Oct 2005 15:51:40 +0100 (CET)

Created on:  MaintDvorak
Type:        porting specific change

Problem Description (Customer terms)
To make the new 'change password' functionality work on BaanIVc on Windows, the Porting Set should tell 3GL whether the user's password will expire soon.

Workaround
n/a

Test Procedure
Run the command line mentioned above.

Affected Executables
badmin.exe




MaintCorelli: # 28083 (BDNT20206): BW msi setup crash when having an 
inconsistent BW msi install

Date: Tue, 25 Oct 2005 14:51:41 +0200 (CEST)

Created on:  MaintDvorak
Type:        bugfix

Problem Description (Customer terms)
When running setup.exe to install new BW version, the setup program generates an Access Violation when there is an inconsistent BW MSI install.
The inconsistent BW MSI install does not contain the file bw.exe, and therefore it is consistent.
(So according Windows this BW is correctly installed, but somehow the file bw.exe is missing).
By running setup.exe of new BW I want to update this inconsistent BW installation so it becomes consistent again.


Workaround
Take care a bw.exe program is there.

Test Procedure
Have a BW MSI installed and remove the bw.exe.
Now run setup.exe of a BW MSI install, in order to upgrade this BW MSI install.
The setup should not crash or stop, but install the new BW version.

Affected Executables
setup.exe




MaintCorelli: # 27929 (BDNT20210): Solve problem in HAL_access_ex function.

Date: Fri, 14 Oct 2005 19:01:26 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
Access check (read + write access) of directory is failing.
E.g. if the directory $BSE_TMP is set to: rwxrwxr-x bsp:bsp 
and the user acohens:tools is member of group bsp.

When doing a touch $BSE_TMP/myfile, with the same permissions, it works.

Workaround
Change directory permissions to 777.

Test Procedure
Refer to Problem Description.

Affected Executables
bshell




MaintCorelli: # 27920 (BDUX20194): MySQL driver: string parameter may be bound 
with length -3

Date: Fri, 14 Oct 2005 13:01:48 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
A query with an expression as follows gives the wrong result

	<string column> & <string column/constant> = <parameter>

	<string column>(<start>;<length>) = <paramater>


Test Procedure
Execute queries with expressions such as described above and verify that the
result is correct.

Affected Executables
mysql_srv6.2




MaintCorelli: # 27901 (BDNT19048): BaanIVc - number of copies in BwPrint not 
taken into account

Date: Thu, 13 Oct 2005 15:53:48 +0200 (CEST)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
The number of copies defined in the driver properties are not considered by BWPRINT. 
In Portingset 6.1c0702 BW73 number of copies was automatically taken from the driver. But 
from portingset version 6.1c0704 till now this seems to be not taken into account anymore.

Workaround
Set the number of copies manually in the printer dialog.

Test Procedure
Set a default printer
Define a default number of copies larger then 1 in the printer settings 
Run BwPrint
Open a bpf file with a number of copies not set (default value is 1)
Choose file->print, a print dialog opens. Check the number of copies, it shoul be the
same as specified in the default printer settings

Affected Executables
bwprint.exe

Motive source
DF:157388




MaintCorelli: # 27898 (BDUX20183): MySQL driver: the maximum length of VARCHAR 
columns is 65533, not 255

Date: Thu, 13 Oct 2005 14:07:05 +0200 (CEST)

Created on:  MaintDvorak
Type:        bugfix

Problem Description (Customer terms)
The MySQL driver restricts the length of string columns (type VARCHAR) to 255,
which is not correct. The maximum VARCHAR column is 65533.

Test Procedure
Create a table with a string column that is large (say 300 or bigger) and 
observe that the column is not split in MySQL.

Affected Executables
mysql_srv6.2




MaintCorelli: # 27894 (BDNT20181): msql: MS-SQL error 4924 should be mapped to 
bdberrno 512

Date: Thu, 13 Oct 2005 13:03:09 +0200 (CEST)

Created on:  DevSchoenberg
Type:        bugfix

Problem Description (Customer terms)
MS-SQL error 4924 should be mapped to bdb_errno 512 (E_BDB_DD_CORRUPT)

Test Procedure
See defect

Affected Executables
msql_srv

Motive source
DF:158924




MaintCorelli: # 27887 (BDUX20136): DF158051: return value of seq.error() etc. 
not correct

Date: Thu, 13 Oct 2005 12:10:44 +0200 (CEST)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
Scopus defect 158051: seq.error(-1) (or any other file descriptor for a
non-existing or already-closed stream) returns 0, instead of -1 as per the
documentation.

Workaround
To know whether a file pointer is valid, the simplest thing is to use

  seq.eof(fp) >= 0

where fp is the file pointer.

Test Procedure
Create a small 3GL script that does seq.error() on
 
 - -1

 - 42

 - a file descriptor that is opened and then closed

and observe that seq.error() returns -1, and sets e to EBADF.

Do the same for

 - a file descriptor that is opened correctly

and observe that seq.error() returns 0, and leaves e unchanged.

Create a small 3GL script that does pipe.error() on

 - the file descriptor returned by pipe.open("echo Hoi", "r")

and observe that pipe.error() returns 0, and leaves e unchanged.

Affected Executables
bshell

Motive source
DF:158051




MaintCorelli: # 27877 (BDUX20178): oracle: error ORA-12899 should be mapped to 
error E_BDB_DD_CORRUPT (512)

Date: Thu, 13 Oct 2005 10:35:50 +0200 (CEST)

Created on:  DevSchoenberg
Type:        bugfix

Problem Description (Customer terms)
Oracle error ORA-12899 "value too large for column %s (actual: %s, maximum: %s)"
should be mapped to bdb_errno 512.

Test Procedure
% bdbpre6.2 -N dbtst120 -x -C 000
<<< Create table dump
% edit the dictionary file. Change
  F
  :NAME:dbtst120.firstnme
  :DOMA:dbs20
into
  F
  :NAME:dbtst120.firstnme
  :DOMA:dbs10
% gcommand gc 000
<<< Using the following file 'gc'
<<< #1 dbtst120	TDRTBL	1
<<< #1 dbtst120	TCRTBL	1
% DBSLOG=01570 bdbpost6.2 -n -x -D .
<<< This import will report error 520. The generated dbs.log file contains the following lines:
	SQL> ora_multi_execute( 0x40083be0 ) do 1, prefetch 1 --> oci_ret -1 
	ora_error 12899 dbs_errno 13899.
	ora_multi_execute( 0x40083be0 ) done - rows 0
	FATAL /vobs/tt/servers/ORACLE_2/ora_native.c:#1739 :

	Error ORA-12899 occurred during execute.
	ORA-12899: value too large for column "B61A"."TDBTST120111"."T$FIRSTNME" (actual: 20, maximum: 10)

	Error BDB-13899 returned.

The new driver generates another log, doing the right mapping:
	SQL> ora_multi_execute( 0x40083be0 ) do 1, prefetch 1 --> oci_ret -1 
	ora_error 12899 dbs_errno 512.
	FATAL /vobs/tt/servers/ORACLE_2/ora_native.c:#1740 :

	Error ORA-12899 occurred during execute.
	ORA-12899: value too large for column "B61A"."TDBTST120111"."T$FIRSTNME" (actual: 20, maximum: 10)

	Error BDB-512 returned.

Affected Executables
ora_srv

Motive source
DF:158868




MaintCorelli: # 27874 (BDUX20090): MySQL driver: incorrect error 109 
(E_BDB_NO_KEY) on corrupt dd

Date: Thu, 13 Oct 2005 10:35:27 +0200 (CEST)

Created on:  MaintDvorak
Type:        bugfix

Problem Description (Customer terms)
bdbreconfig may incorrectly report error code 109 (E_BDB_NO_KEY) if an attempt
is made to drop a non existing column.

Test Procedure
Using bdbreconfig try to drop the same column twice. The first time it should go
without problems, the second time error 512 (E_BDB_DD_CORRUPT) should be 
reported.

Affected Executables
mysql_srv6.2




MaintCorelli: # 27815 (BDUX19699): no outpt written by run.(baan.)prog

Date: Tue, 11 Oct 2005 09:52:15 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
Can not redirect standard output and standard error of a program started with 
run.prog and run.baan.prog.

Workaround
Use shell function

Test Procedure
Run the next script and check, if there is readable text in the files. 

function main() {
  long retval
  retval=run.baan.prog( "bshell", "-v", RP_WAIT, ""
			, "baan.stdout.log",	"baan.stderr.log" ) 
  retval=run.prog( "ls", "", RP_WAIT, ""
			, "stdout.log", "stderr.log" )
}

Check, if the files have been created and if there's readable text in the files.

Affected Executables
sh_server6.2

Motive source
DF:226814




MaintCorelli: # 27781 (BDUX20107): replace PATH_MAX by MAX_PATH

Date: Thu, 6 Oct 2005 17:32:50 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
on iSeries compress6.2 fails when filename length > 64
bshell function dir.open() and dir.open.tree() may also fail

Workaround
use only filenames with length < 40

Test Procedure
BSE_SHLIB=/home/PORT/directory_with_a_very_long_name/shlib /home/PORT/directory_with_a_very_long_name/bin/compress6.2 -f /home/PORT/directory_with_a_very_long_name/porting/PA.9876/portingset/OS4005.1/OS4005.1.bpk

Affected Executables
bshell
compress




MaintCorelli: # 27761 (BDUX20088): MySQL: Wrong log message: Illegal idxopt 12 
with type 0 encountered

Date: Thu, 6 Oct 2005 10:54:56 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
The following message is logged in $BSE/log/log.mysql:

	Illegal idxopt 12 with type 0 encountered

Test Procedure
Run a query on any table and observe that the log message is not generated
anymore.

Affected Executables
mysql_srv6.2




MaintCorelli: # 27757 (BDNT19813): Mouse wheel support in BW grid

Date: Wed, 5 Oct 2005 17:20:57 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
Currently the mouse-wheel to scroll through a BW grid only works when you have a special
mouse-driver supporting the mouse-wheel installed.

Workaround
Install specific mouse driver.

Test Procedure
For example, start ttaad2500m000 User Data,
and make the window small, to get a vertical and horizontal scrollbar.
When turning the mouse wheel now, you should scroll vertical.
When holding the shift-key and turning the mouse wheel, you should scroll horizontal.

Affected Executables
bw.exe

Motive source
DF:159698




MaintCorelli: # 27752 (BDUX20082): Oracle: removed obsolete resource

Date: Wed, 5 Oct 2005 16:42:27 +0200 (CEST)

Created on:  DevSchoenberg
Type:        porting specific change

Problem Description (Customer terms)
The Oracle "alter session set optimizer_goal = XXX" sysntax is desupported since Oracle 9i.
This implies that the resource optimizer_goal has no use anymore.

Affected Executables
ora_srv




MaintCorelli: # 27732 (BDUX19221): Get MaintCorelli to compile on z/OS

Date: Tue, 4 Oct 2005 14:16:08 +0200 (CEST)

Created on:  MaintCorelli
Type:        porting specific change

Problem Description (Customer terms)
Port to z/OS 1.6 / DB2 V8.

Workaround
n/a

Test Procedure
n/a

Affected Executables
all




MaintCorelli: # 27720 (BDUX20044): MySQL driver: give error 512 on missing 
column in table

Date: Mon, 3 Oct 2005 16:14:25 +0200 (CEST)

Created on:  MaintDvorak
Type:        bugfix

Problem Description (Customer terms)
If the datadictionary does not match with the table in the database the driver
should return error code 512 (E_BDB_CORRUPT), but it returns the error codes
of MySQL bumped by 1000.

Also, if an attempt to lock a row fails because the table is locked, it is
good enough to return E_BDB_LOCKED.


Test Procedure
Modify the dd file of an existing table by adding a column to it. Now select all
colums (using *) of the table and observe that error 512 is returned.

Affected Executables
mysql_srv6.2




MaintCorelli: # 27696 (BDNT20007): intern compile problem sourcecode protection

Date: Fri, 30 Sep 2005 11:03:46 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
bic_pre also needed to link in biclib and the slm api




MaintCorelli: # 27669 (BDUX19871): Backport of MySQL driver to Corelli

Date: Thu, 29 Sep 2005 10:48:40 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
A backport of the MySQL driver to 7.1d is needed by Chicago Metallic.

Test Procedure
It should be possible to use the MySQL driver as any other database driver.
You will need a (very) recent version of MySQL to test against (5.0.14 has still 
some problems, 5.0.16 is the targeted GA version).

Affected Executables
mysql_srv6.2
mysql_maint6.2




MaintCorelli: # 27647 (BDNT19979): TCS800-148369 BW Installer should not update 
full install

Date: Wed, 28 Sep 2005 15:31:24 +0200 (CEST)

Created on:  MaintDvorak
Type:        bugfix

Problem Description (Customer terms)
The installer recognizes an AS installation (Logic and BW) on the local machine 
and wants to overwrite it.

It should skip the AS installation because this is an other product.


Test Procedure
See problem description

Affected Executables
setup.exe

Motive source
TCS:800-148369




MaintCorelli: # 27637 (BDUX19969): creat.tmp.file: handle empty pathname 
correctly.

Date: Wed, 28 Sep 2005 12:31:18 +0200 (CEST)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
The bshell function creat.tmp.file$() should generate a temporary file-name,
create that file and return the file-name.
But in the bshell's from the nightly builds from 2005/09/21 this function returns 
an empty string as file-name and doesn't create the file.

Workaround
Specify current working directory as path.

Test Procedure
Execute the following 3GL program:

function main()
{
	string	name(512)
	name = creat.tmp.file$()
}

Now "name" should be filled with the generated tmp filename.

Affected Executables
bshell




MaintCorelli: # 27619 (BDNT19952): BW.EXE compilation error fixed

Date: Tue, 27 Sep 2005 16:36:16 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
BW.EXE build fails.

Workaround
n/a

Test Procedure
see BDNT_19844

Affected Executables
bw.exe
bwc.dll




MaintCorelli: # 27614 (BDUX19940): HPUX trusted security: do not allow to log in 
when account is locked out

Date: Tue, 27 Sep 2005 14:49:09 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
Login with BaanLogin protocol granted, although unix os account is locked.

Workaround
Take care that the Baan user is disabled or removed in case he is not allowed to log in.

Test Procedure
On an HPUX system running in Trusted Security mode:
Take care that the os account is locked out.
(Administrative lock-out or locked out because of too many unsuccessful login attempts).
Log in Baan using BW and BaanLogin protocol selected.
Now it should not possible to log in.
The following error is generated:

1: Error 4 (-1013: User account/profile is disabled.) : baanlogin failed host 'mega' username 'trainer' command...
2: Error : bw failed to connect to mega!bshell
or

1: Error 4 (-1018: User account/profile is disabled because of too many unsuccessful login attempts.) : baanlogin failed host 'mega' username...
2: Error : bw failed to connect to mega!bshell

Affected Executables
blogind6.x

Motive source
DF:157630




MaintCorelli: # 27567 (BDUX19926): intern Source code protection more secure 
Part II

Date: Mon, 26 Sep 2005 09:44:16 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
sizeof() without constant doesn't compile on the giotto.




MaintCorelli: # 27534 (BDNT19844): Port number for BaanLogon protocol increased

Date: Fri, 23 Sep 2005 10:34:30 +0200 (CEST)

Created on:  MaintMahler
Type:        bugfix

Problem Description (Customer terms)
Port number for BaanLogon protocol is limited to 10000.

Workaround
n/a

Test Procedure
configure BaanLogon protocol for a number greater 10000 and save it.

Affected Executables
bw.exe




MaintCorelli: # 27520 (BDNT19781): Virtual Memory leak when ora_timeout is set.

Date: Thu, 22 Sep 2005 16:53:01 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
When ora_timeout is set the total virtual memory of the Oracle driver only increases after each update.
For example, when running the Export from session tgbrg125m000, using a small selection, the virtual memory of the
Oracle driver process increases with about 35MB per run.
In case of the customer, who is using bigger data set's, the virtual memory of the Oracle driver process hits the
Windows OS limit of 2GB.

Workaround
Do not set ora_timeout.

Test Procedure
On Windows, running Baan and Oracle:
Log in Baan, and start ttaad4100
Start the Performance Monitor, and add a counter.
Select Process.
Select the ora8_srv.exe process.
Select counter Virtual Bytes.
Now, do an update and watch the Virtual Bytes.
The number of Virtual Bytes should not be increased after the update.


Affected Executables
ora8_srv.exe

Motive source
DF:159137




MaintCorelli: # 27512 (BDUX19878): In repgen, add line directive for print 
expressions for better compile messages.

Date: Thu, 22 Sep 2005 08:34:54 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
When compiling a report, errors can occur in Output Conditions on Details
or in Print Expressions on fields. These errors indicated in the intermediate
source generated by repgen.
For a developer, it is hard to find out where exactly the error is caused. He
has to check manually all Output Conditions and Print Expressions to find
the piece of code that is modelled incorrectly.

Workaround
Search manually for the cause of the error.

Test Procedure
Use an undeclared variable in a Output Condition or a Print Expression and then
compile the report. The error message should be extended with the context.

Affected Executables
repgen6.2
bic6.2

Motive source
DF:155848




MaintCorelli: # 27511 (BDUX19875): intern Source code protection more secure

Date: Thu, 22 Sep 2005 08:34:02 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
If customer deletes tab.reg he can work unsecure.
In this project code has been added so that customer also need to remove tt security file to work unsecured.
In al_secu.c process.h was included outside #ifdef __WIN32__, this gave compilation problems.




MaintCorelli: # 27502 (BDNT19870): Repair merge error (BDNT19859)

Date: Wed, 21 Sep 2005 13:49:48 +0200 (CEST)

Created on:  DevSchoenberg
Type:        bugfix

Problem Description (Customer terms)
1) The merge of BDNT19859 modifications from Reger to MaintCorelli, MaintMahler,  MaintDvorak and DevSchoenberg was not correct.
This is causing build problems.

2) New argument of HAL_access_ex() wrongly declared.


Test Procedure
Building libhal.lib and albsh.lib should not fail anymore.

Affected Executables
bshell




MaintCorelli: # 27494 (BDUX19583): intern Source code protection

Date: Wed, 21 Sep 2005 12:26:06 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
No license is needed to compile a source




MaintCorelli: # 27483 (BDNT19859): Creation of tmp file fails because of EACCES 
error (continuation of BDNT19423)

Date: Wed, 21 Sep 2005 08:59:28 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
After project BDNT19423: In case you do not have write-access to the directory (causing EACCESS error),
function creat.tmp.file$() will try again in an endless loop.

Workaround
Make sure that the user has read and write access to directory.

Test Procedure
Refer to project BDNT19423.

Affected Executables
Bshell

Motive source
DF:156196




MaintCorelli: # 27480 (BDNT19855): Merge project BDNT19800 to Corelli

Date: Tue, 20 Sep 2005 17:03:11 +0200 (CEST)

Created on:  MaintCorelli
Type:        porting specific change

Problem Description (Customer terms)
Keep acc.c in sync between versions

Workaround
NA

Test Procedure
Use "acc -acc ebcdic" for instance (see SLM build)

Affected Executables
acc on AS/400




MaintCorelli: # 27436 (BDUX19816): Solve build problem after BDNT17466

Date: Mon, 19 Sep 2005 14:59:53 +0200 (CEST)

Created on:  MaintDvorak
Type:        bugfix

Problem Description (Customer terms)
After project BDNT17466 and BDNT19782, taudit6.2 did not build anymore.

Test Procedure
Make a build of porting set.

Affected Executables
bshell, etc




MaintCorelli: # 27387 (BDNT17466): Disable warning about attempt to start more 
than one dll

Date: Fri, 16 Sep 2005 09:18:09 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
The Windows Event log is filled up with the following error:

Env	 B50C (c:\baan)
Prog	 SOCKET<baan@NTBAAN:1684>) file \general\lib\nw_1\IPC_GLOB.C # 903
Keyword	 ipc_info
Username	 lrizzo	type S	language 6
Process	 0xa764
ProcessSet	 baan@NTBAAN:1684
Lasterror	 0
Errno	 0
bdberrno	 0
Message	 Warning: Attempt to start more than one shared library: Trying to fallback to standalone driver with same logical name (see %s/lib/ipc_info)

Therefore it is troublesome to see if there are any (really) serious errors.


Workaround
Ignore error.

Test Procedure
On a Windows system running Baan on MS SQL Server:
Have the for example the following lines in %BSE%\lib\tabledef6.2:

ttaad200:000:msql7(MSQL_SERVERHOST=NLBAWKVDDOOL1,MSQL_LEVEL1=0):N
*:*:msql7(MSQL_LEVEL1=0,MSQL_SERVERHOST=NLBAWKVDDOOL1):N

you see the parameters are the same, but in a different order.

The %BSE%\lib\ipc_info should contain the default lines, like:

msql7           d       msql7_srv
msql7           p       ${BSE}\bin\msql7_srv.exe

you see the first line is for combo (d), and the second line is normal (p).

Log in Baan.
Now a ntbshell.exe process will be started which runs in combo  mode.
(So there is only a ntbshell.exe process and no seperate msql7_srv.exe process, because 
the bshell runs in combo mode).

Open session ttaad2500m000 (Maintain User Data).
In this way you will access table ttaad200 of company 000, and a second database driver will be started.
Now there is also a msql7_srv.exe process.
And the following warning is logged in the Windows Event Log:

Message	 Warning: Attempt to start more than one shared library: Trying to fallback to standalone driver with same logical name (see <BSE path>/lib/ipc_info).

Now, add the following line to the file %BSE%\lib\defaults\all:

hide_combo_errors:1

Logout of Baan and Log in Baan again, and repeat the steps.
Now no warning should be logged anymore.

Affected Executables
bshell

Motive source
DF:158766




MaintCorelli: # 27338 (BDNT19736): DF158932 Do not sent TAB codes to Bshell in 
an Edit field

Date: Wed, 14 Sep 2005 16:41:50 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
When copy/pasting text from an Excell Worksheet, and the cell format is a merged cell,
TAB characters are appended to the end of the inserted data.

Test Procedure
See problem description.

Affected Executables
bw.exe

Motive source
DF:158932




MaintCorelli: # 27309 (BDUX19716): Oracle driver: FIRST_ROWS(n) hint

Date: Tue, 13 Sep 2005 17:01:36 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
Oracle driver does generate the old style FIRST_ROWS hint instead of the FIRST_ROWS(N) hint

Test Procedure
PBC Benchmark
regression tests

Affected Executables
ora8_srv

Motive source
DF:158990




MaintCorelli: # 27272 (BDUX19676): Merge from BDUX19674

Date: Mon, 12 Sep 2005 12:58:30 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
Compilation errors.




MaintCorelli: # 27260 (BDUX19565): Error 305 occurs on random record

Date: Fri, 9 Sep 2005 14:49:49 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
Error 305 (wrong row returned) occurs during run of parallel jobs on AS/400.

Workaround
Set the next parameter in $BSE/lib/defaults/db_resource:

db4_max_ret_rows:1

Test Procedure
It is not possible to test this. This error only occurs in a specific situation.

Affected Executables
db4_srv6.2

Motive source
DF:150930




MaintCorelli: # 27254 (BDNT19656): Informix RDBMS includes added for Windows

Date: Fri, 9 Sep 2005 13:56:23 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
Compiler warnings for serv_table on windows.




MaintCorelli: # 27227 (BDUX19631): Introduced autocommit for db.clear.table 
function

Date: Thu, 8 Sep 2005 14:03:25 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
Bshell function db.clear.table(tid, NO.ROLLBACK) does not work correct for Informix drivers.


Workaround
add more logical logs to your database in case you are getting error -458 (long 
transaction abort)

Test Procedure
1.) add to db_resource

    use_cleartable_autocommit:1

2.) run a gcommand script to clear a table with NO.ROLLBACK.
    the number of rows in table should be > 1000

    e.g. #1 ttaad050	TCLTBL	1
         #1 ttaad050	DCOM	0

Affected Executables
inf_srv6.2




MaintCorelli: # 27191 (BDUX19586): Startup sessions do not work

Date: Tue, 6 Sep 2005 11:18:53 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
Startup sessions don't work on AS/400.

Test Procedure
Define a startup session for a user and test, if the session starts. 

Affected Executables
ipc_boot6.x

Motive source
DF:157435




MaintCorelli: # 27160 (BDNT19423): Creation of tmp file (creat.tmp.file$()) 
fails because of EACCES error

Date: Fri, 2 Sep 2005 16:56:18 +0200 (CEST)

Created on:  MaintReger
Type:        bugfix

Problem Description (Customer terms)
If I want to print a large range of data, I get the error message 'Could not open temporary file'. After selecting the 'OK' button, the error message 'Could not copy text to temporary file'. No report is generated.

In the event viewer, the following message is displayed:
Env	 b52a01 (e:\baan\b52a01)
Prog	 bshell file \general\lib\al_1\AL_FPATH.C # 206
Keyword	 sopen
Username	 beda	type N	language 2
Process	 0xff00
ProcessSet	 BEDA@BGSCLT7M:2552
Lasterror	 5 (Access is denied.)
Errno	 13 (Permission denied)
bdberrno	 0
Message	 
Error during sopen(f:/baan/b52a01/tmp/tmp016776929, f:/baan/b52a01/tmp/tmp016776929)

Printing a smaller range of data doesn't cause this problem.

When debugging the 4GL program code, the problem does not appear.
When turning on Bshell tracing, like -dbgfdev or -dbgflow, the problem does not appear.
It seems when delaying the executing by debugging or tracing, the problem does not occur anymore.
Therefore it is very hard to investigate this issue.

Workaround
Print smaller reports (smaller selections).

Test Procedure
Because this is a timing issue, it is hard to give a test procedure for this problem.
I tested it on the customer system, where the problem could be reproduced consistently by running a customized printsession.
Further I tested if the current behavior was not changed.
Therefore tested again the test procedure of VHNT3087.
And did the following test:

C:\baan\baan5c\bin>bic_explode otdsls4100
c:/baan/baan5c/application/tdB50U_c_stnd/otdsls/osls4100

you see the object is found.
Now change the file permissions of this object in that way that you are not allowed to read this file anymore, and
do again the explode:

C:\baan\baan5c\bin>bic_explode otdsls4100
bic_explode:
Error during sopen(c:/baan/baan5c/application/tdB50U_c_stnd/otdsls/osls4100, c:/baan/baan5c/application/tdB50U
_c_stnd/otdsls/osls4100)
c:/baan/baan5c/application/tdB50_c_sp02/otdsls/osls4100

now you see an error is logged. This is correct.


Affected Executables
bshell

Motive source
DF:156196




MaintCorelli: # 27068 (BDUX19497): Multi-byte substring operator gives out of 
dims, but should not (backport of BDUX12733

Date: Mon, 29 Aug 2005 16:22:45 +0200 (CEST)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
Scopus *case* 2055357: When getting a substring from a multi-byte string in the
BaanC language, sometimes an 'index out of dims' occurs.

Test Procedure
Run the following simple 3GL program:
   function main()
   {
        string s1(10)
        string s2(10) mb
        string substr(100)

        s1 = "Hallo67890"
        s2 = "Hallo67890"

        substr = s1(8;5)

        substr = s2(8;5)
   }

It should not give an 'index out of dims' error.

Affected Executables
bshell




MaintCorelli: # 26890 (BDNT19344): TCS800-148085 Update InstallShield 3 based BW 
takes a long time

Date: Thu, 18 Aug 2005 15:12:19 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
When upgrading Baan Windows from an InstallShield 3 version of BW,
Un-installing the InstallShield 3 version sometimes takes a long time.
During this un-install action, nothing seems to happen and no window is visible.

Test Procedure
See problem description.

Affected Executables
setup.exe

Motive source
TCS:800-148085




MaintCorelli: # 26881 (BDUX19342): DF158444: Fix wrong LINE statements when 
adding bitsets

Date: Thu, 18 Aug 2005 11:16:00 +0200 (CEST)

Created on:  MaintMahler
Type:        bugfix

Problem Description (Customer terms)
Scopus defect 15844: compiling code such as the following (comments are not
relevant):

	domain ttadv.tuse x                            | line 42
	x = ttadv.tuse.read + ttadv.tuse.delete        | line 43
	some-other-statement                           | line 44 (wrong: 43)
	more statements                                | line 45 (wrong: 44)

causes lines 44 and later to be numbered wrongly (in this case: they are
numbered as 43, 44, ...), when the script is compiled in debug mode.  This
causes the debugger to point to the wrong 'current line', which is very
confusing.

Workaround
Call the function bit.or() to 'add' the bits of the bitset together.

Test Procedure
See problem description.

Affected Executables
bic

Motive source
DF:158444




MaintCorelli: # 26845 (BDUX19282): Make yy lib bugwards compatible again

Date: Tue, 16 Aug 2005 12:15:24 +0200 (CEST)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
Make expressions behave backwards compatible in case
of range errors like [_-.] and [z-a]

Workaround
Fix the faulty expression

Affected Executables
all




MaintCorelli: # 26809 (BDUX19278): Problem staring audit with modified tabledef 
file

Date: Fri, 12 Aug 2005 15:55:14 +0200 (CEST)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
Audit server could not be started.

Workaround
never (!) change the tabledef file by hand

Test Procedure
ksh % BDB_DRIVER="oracle8 ()" qptool6.1 -c 0 -q "select * from dbtst120"
results in error 505. (look at space after driver name).
The changed porting set should accept this command.

Affected Executables
bshell
qptool

Motive source
DF:158391, case 1155221




MaintCorelli: # 26709 (BDUX19140): S226046: Query with tables from multiple 
companies in multiple db dumps core

Date: Wed, 3 Aug 2005 16:14:01 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
If a query contains an OR and each branch contains a subquery and both 
subqueries cannot be handled by the same database driver, the client (qptool or 
bshell) crashes.

Test Procedure
Set-up a scenario as described in the Problem Description (Technical terms). 
Execute the query and observe that no crash occurs and that the proper result is 
returned.

Affected Executables
bshell
qptool

Motive source
DF:226046



MaintCorelli: # 26625 (BDNT19106): Show new version in BW setup update dialog

Date: Thu, 28 Jul 2005 12:26:34 +0200 (CEST)

Created on:  MaintDvorak
Type:        bugfix

Problem Description (Customer terms)
New Baan Windows version not visible when existing version is updated.

Test Procedure
Update installed Baan Windows version using MSI setup.

Affected Executables
setup.exe




MaintCorelli: # 26612 (BDUX19090): error logging improvement

Date: Wed, 27 Jul 2005 14:40:16 +0200 (CEST)

Created on:  MaintDvorak
Type:        bugfix

Problem Description (Customer terms)
Unclear error messages about wrong range expressions in domains.

Affected Executables
all

Motive source
TCS:800-147700



MaintCorelli: # 26541 (BDUX19025): S157057: QP: Incorrect result for re-executed 
distributed subquery

Date: Tue, 19 Jul 2005 14:08:03 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
If a query containing a subquery that contains a _compnr = <parameter> expression
is executed multiple times with different values for the parameter, and for these
different values different the subquery must be handled by different database
drivers, then the result of the query may be incorrect.


Workaround
Use a nested query to filter out the rows that should be filtered by the subquery.

Test Procedure
Set up tabledef as follows (be sure not to have set BDB_DRIVER):

	dbtst100:100:oracle8(SITE=2)
	*:*:oracle8

>From the command line enter the following:

	qptool -a pefbefbefbc -c 0 -q "
		select empno, sex
		from dbtst120
		where empno inrange 10 and 100
			and sex <> :1
		and not exists (
			select deptno
			from dbtst100
			where deptno = workdept
				and _compnr = :1
		)
		order by empno
	"

Enter the following types/values when requested for:

	L
	1
	L
	0
	L
	1

Verify that you get the correct result, and that in fact the "not exists" filter
operates correctly every time.

Affected Executables
bshell
qptool

Motive source
DF:157057




MaintCorelli: # 26536 (BDUX19021): Small error with TMS profiling

Date: Tue, 19 Jul 2005 12:00:58 +0200 (CEST)

Created on:  DevSchoenberg
Type:        bugfix

Problem Description (Customer terms)
TMS (times.on/off) level not correctly displayed in Call Graph Profiler

Test Procedure
Run Call Graph Profiler with times.on/off functions in test object

Affected Executables
bshell




MaintCorelli: # 26489 (BDUX18990): Core dump when 3th argument of xmlattribute 
is not specified

Date: Fri, 15 Jul 2005 11:19:34 +0200 (CEST)

Created on:  MaintDvorak
Type:        bugfix

Problem Description (Customer terms)
ERROR
Core dump when calling xmlAttribute$ for a non-existing attribute without 
specifying a 3rd argument (default value).

REPRODUCTION
function main()
{
	long	node
	string	attr$(10)

	node = xmlNewNode("TEST")
	attr$ = xmlAttribute$(node, "ATTR", "")	| OK
	attr$ = xmlAttribute$(node, "ATTR")	| Core dump
}

Programmer's Guide mentions:
When default.value is not specified, an empty string is used.

NOTES
- This error is encountered on NT (nlbawptg3), not on sprinter (AIX)
- Also check other xml function that returns strings:
	* xmlAttributeName$()
	* xmlData$()
	* xmlDataElement$()
	* xmlName$()


Workaround
Specify a default as third argument.

Test Procedure
compile and run the script under NT

Affected Executables
bshell6.2

Motive source
TCS:800-146644




MaintCorelli: # 26484 (BDNT18861): MSQL level 1 driver corrupts table if array 
interface is used

Date: Fri, 15 Jul 2005 10:08:19 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
Tables get compeltely corrupted, when reorganizing them in the following conditions:
- SQL Server level 1 database is used.
- Array interface is enabled (msql_max_arrsz:3)


Workaround
don't use array interface or set it to size 1

Affected Executables
msql7_Srv

Motive source
DF:156146




MaintCorelli: # 26464 (BDNT18974): Port is logged in network order

Date: Thu, 14 Jul 2005 09:42:17 +0200 (CEST)

Created on:  MaintCorelli
Type:        bugfix

Problem Description (Customer terms)
Portnumber 512 is logged as port 2.




MaintCorelli: # 26457 (BDUX18964): JVMI: proper log message

Date: Wed, 13 Jul 2005 15:08:39 +0200 (CEST)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
No logging when bshell can not start the Java Virtual Machine.

Test Procedure
start tmboaserver without license.

Affected Executables
bshell

Motive source
DF:156983




MaintCorelli: # 26381 (BDUX18884): Repgen: suppr__ strings are not MB

Date: Thu, 7 Jul 2005 14:41:35 +0200 (CEST)

Created on:  MaintDvorak
Type:        bugfix

Problem Description (Customer terms)
Suppressing of MB fields in reports is not working correctly.
Some MB fields are cut off during printing.

Test Procedure
run whinh3412m100 with all lines and containing MB data.

Affected Executables
repgen6.2

Motive source
TCS:200-105888




MaintCorelli: # 26377 (BDUX18888): Oracle shared libraries not found

Date: Thu, 7 Jul 2005 12:20:34 +0200 (CEST)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
Unable to use ora_inst6.1 to install a baan envrionment, using latest porting 
set 6.1c06.08. 

Workaround
In environment set the next parameter depending on the operating system:

# HP-UX
SHLIB_PATH=$ORACLE_HOME/lib32

# AIX
LIBPATH=$ORACLE_HOME/lib32

# Sys V rel 4
LD_LIBRARY_PATH=$ORACLE_HOME/lib32

Test Procedure
Use ora8_inst6.x to do some database actions.

Affected Executables
ora_env6.x

Motive source
DF:157574




MaintCorelli: # 26373 (BDUX18903): Oracle driver: gcommand DRIDX does not work

Date: Wed, 6 Jul 2005 17:15:08 +0200 (CEST)

Created on:  MaintMahler
Type:        bugfix

Problem Description (Customer terms)
gcommand command TDRIDX does not work on Oracle.

Workaround
bdbreconfig -Z (reorganize table)

Test Procedure
create a gcomman file:
	#1 dbtst120 TDRIDX 2
	#1 dbtst120 TCRIDX 2

run DBSLOG=01570 gcommand <comman file> 0
verify in the dbs.log file that the index is dropped and recreated in Oracle.

Affected Executables
ora8_srv




MaintCorelli: # 26366 (BDNT18899): BwPrint: line drawing on fax

Date: Wed, 6 Jul 2005 17:14:22 +0200 (CEST)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
BwPrint prints vertical lines incorrect to LanFax printer

Test Procedure
See Scopus defect

Affected Executables
bwprint.exe

Motive source
DF:156605




MaintCorelli: # 26317 (BDNT18047): Installer crash after selecting local 
SQ2L2000 instance

Date: Mon, 4 Jul 2005 17:02:13 +0200 (CEST)

Created on:  MaintBaanIVc
Type:        bugfix

Problem Description (Customer terms)
CD is BD0024.2731, poritingset used is 6.1c.07.06 labeled SSA Baan IVc4 for Microsoft Windows Master CD US English BD0024.2968

When following the installation, the installer crashes and the following error appears in the event viewer:
Event Type:	Error
Event Source:	Application Error
Event Category:	(100)
Event ID:	1000
Date:		6/9/2005
Time:		3:25:42 PM
User:		N/A
Computer:	TESTSERVER
Description:
Faulting application _INS0432._MP, version 3.0.116.0, faulting module oleaut32.dll, version 5.2.3790.0, fault address 0x0001cac7.

For more information, see Help and Support Center at http://go.microsoft.com/fwlink/events.asp.


Workaround
The sQL Server should be configured with Binary sort order.

Test Procedure
Have SQL Server configured with sort order:

case-insensitive, accent-sensitive, kanatype-insensitive, width-insensitive for Unicode Data, SQL Server Sort Order 52 on Code Page 1252 for non-Unicode Data

and run the installer.
It should not crash, but give you the following error message:

'SQL Server is not configured with binary sort order. Please use the 'Rebuild Master Database' option in SQL Server
setup to change SQL Server's sort order to 'binary' and restart the Baan installation.
Exiting Setup with an error."

Affected Executables
msql7adm.dll

Motive source
DF:157363