SQLBase 12.0 SP1 - SQL-Error 10446 Unicode conversion failed

[Chacay]

Hi,

we can't select a long varchar from a SQLBase Server 12.0 SP1 running under Linux/CentOs7. If the long varchar column contains only ASCII characters, there will be no problem. If the long varchar contains a special character, i.e. 0x5F, the SQL-Error 10446 "Unicode conversion failed" happens. I don't know which other characters above 0x7F will force this error too.

We have found out, that when we compile (TD 6.2 SP5) our testprogram with target ".NET WPD Desktop (EXE)", the error never happens. As "Win32 Target Type", we can't select under Linux columns with the mentioned character.

If we connect with the Win32-exe-Version to SQLBase 12.0.1 10862 running under Windows, the error never happens. Only with the Linux server we have the problem. Ah, if we connect with SQLTalk to the Linux server, we can select all the data without errors!

Code: Select all

    connected with | SQLBase Server 12.0 SP1 | SQLBase Server 12.0.1 10862 |
                   | running under Linux     | running under               |
Testing with       | CentOS 7.0              | Windows Server 2012 R2      |
----------------------------------------------------------------------------
SQLTalk.exe        | no error                | no error                    |
compiled for .NET  | no error                | no error                    |
compiled for Win32 | error 10446             | no error                    |

We have no idea, what is going wrong. Will there be a workaround?

I have attached a democase to show what happens. The file kf.unl contains the table with the test pattern and the file linuxtest.apt a short application.

Thanks for any help.

Joachim

[Mike Vandine]

Sorry for the late reply.

It's certainly not that we are not interested in the product or problems. Our support priorities have always been for customers with a technical support contract and then trying to help customers on the forum. This has NOT changed since OpenText aquired us.

I will try to get one of the Developers to have a look at this. I don't have the environment to do so. However, be advised that we are trying to get SB12.0.2 out the door in the next couple of days, so a response from them might be a bit slow.

Best regards,

Mike Vandine

[Chacay]

Hi Mike,

thank you for your help and interim report. I hope there will be a solution. We look forward to a response from the developers.

Best regards,

Joachim

[dosborne]

Be sure to start the server on linux with the proper LANG setting.
In my environment, I started it with LANG=en_US.iso8859-1, and was able to fetch the data correctly. With the default of en_US.UTF-8, I got the same error.

If you are not using NCHAR columns, then the server and client code pages must match.

For example:

Code: Select all

LANG=en_US.iso8859-1 LD_LIBRARY_PATH=. ./dblxsrv ini=sql.ini

[Mike Vandine]

Thanks, Daniel!!

Best regards,

[Chacay]

Thanks a lot, Daniel and Mike!

Not only the test case, but also our entire ERP system runs very well. Okay, we have a lot of work ahead of us to test all the functionality, but it looks very good.

One last question, sorry. Does the server code page change affect any sort order or the interpretation of dates or numbers?

Best regards,

Joachim

[Mike Vandine]

Hi Joachim,

Dates should not be affected at all.

Cheers!