B
Boris
I finally managed to create a small testcase to reproduce a bug somewhere in
Access, ADO, COM etc. - see my original posting in October 2004:
http://groups.google.com/groups?lr=&[email protected]&rnum=1
As I don't know where the bug is located I crosspost this message to several
newsgroups as I might need help from different experts to track down the
bug.
The testcase is an Access 2003 add-in developed in C# (one file with about
150 lines of code). When you open any database and start the testcase
(you'll find a new button in the menu bar) it will establish a connection to
the loaded database, read all data of the first table found, delete the
first row and write all data back to the database. If this is done in a
thread (the writing all data back part) an exception will be thrown which
doesn't seem to fit at all: "Selected collating sequence not supported by
the operating system".
When you search for this error on the Internet you'll find some articles in
the Microsoft Knowledge Base that describe other possible reasons. And you
will find some postings by confused people who apparently got this error
message just like me in a situation where it doesn't make any sense.
The most weird thing now is how to make the add-in work perfectly: When you
start Access and open a database just doubleblick on any form, view, table
etc. to open it. If you start the add-in then no exception will be thrown.
When opening an Access object manually something seems to be done which
makes it possible for the add-in later to update the database in a thread.
But what is it? What do you have to do in your Access add-in to make it work
without refactoring and getting rid of the thread or telling all your users
to first open any object in the database window and then starting the Access
add-in (which is kind of ridiculous of course)? Or did I run into some COM
thread stuff which isn't a bug at all but just normal in the COM world?
I would be happy to track down this bug and find a fix to include in the
Access add-in. I uploaded five files to
http://www.highscore.de/testcase/accessaddin.html: A compiled version and
the source code of the testcase, a database to play around with (you can use
any database but the Access add-in will try to delete a row in the first
table) and two files to register and unregister the add-in. I use Microsoft
Access 2003 and .NET 1.1 and can reproduce the bug. My Windows XP and Office
2003 version have all patches installed - the only exception is .NET
Framework 1.1 which hasn't the service pack 1 installed.
If you have time please download the files from
http://www.highscore.de/testcase/accessaddin.html and follow the description
to reproduce the bug. I would like to know if others can reproduce this bug.
If anyone has an idea what is going wrong here I'd be happy to know this,
too!
Boris
Access, ADO, COM etc. - see my original posting in October 2004:
http://groups.google.com/groups?lr=&[email protected]&rnum=1
As I don't know where the bug is located I crosspost this message to several
newsgroups as I might need help from different experts to track down the
bug.
The testcase is an Access 2003 add-in developed in C# (one file with about
150 lines of code). When you open any database and start the testcase
(you'll find a new button in the menu bar) it will establish a connection to
the loaded database, read all data of the first table found, delete the
first row and write all data back to the database. If this is done in a
thread (the writing all data back part) an exception will be thrown which
doesn't seem to fit at all: "Selected collating sequence not supported by
the operating system".
When you search for this error on the Internet you'll find some articles in
the Microsoft Knowledge Base that describe other possible reasons. And you
will find some postings by confused people who apparently got this error
message just like me in a situation where it doesn't make any sense.
The most weird thing now is how to make the add-in work perfectly: When you
start Access and open a database just doubleblick on any form, view, table
etc. to open it. If you start the add-in then no exception will be thrown.
When opening an Access object manually something seems to be done which
makes it possible for the add-in later to update the database in a thread.
But what is it? What do you have to do in your Access add-in to make it work
without refactoring and getting rid of the thread or telling all your users
to first open any object in the database window and then starting the Access
add-in (which is kind of ridiculous of course)? Or did I run into some COM
thread stuff which isn't a bug at all but just normal in the COM world?
I would be happy to track down this bug and find a fix to include in the
Access add-in. I uploaded five files to
http://www.highscore.de/testcase/accessaddin.html: A compiled version and
the source code of the testcase, a database to play around with (you can use
any database but the Access add-in will try to delete a row in the first
table) and two files to register and unregister the add-in. I use Microsoft
Access 2003 and .NET 1.1 and can reproduce the bug. My Windows XP and Office
2003 version have all patches installed - the only exception is .NET
Framework 1.1 which hasn't the service pack 1 installed.
If you have time please download the files from
http://www.highscore.de/testcase/accessaddin.html and follow the description
to reproduce the bug. I would like to know if others can reproduce this bug.
If anyone has an idea what is going wrong here I'd be happy to know this,
too!
Boris