> I might be wrong on this - this might have been done after B5 was
> released. I'm making so many changes in the EDB Manager, it's hard to
> keep up.

I'm glad it's in there for the next release then!  


--
David Cornelius
CorneliusConcepts.com

Tim Young [Elevate Software] wrote:

><< Hopefully someone has, since there is a driver for it in DA, but noone
>at RO seems to know either! There was me thinking if I waited while EDB
>was around a while, the problems would be less >>
>
>We didn't write the driver for DA, so I have no idea who did, but I would
>assume that RO did so.

They did, but it looks like it was for version 1, but they've said they're
going to see about getting a copy of version 2 (I guess no-one else is
using it yet).


--
Jeremy Knowles

Jeremy,

<< They did, but it looks like it was for version 1, but they've said
they're going to see about getting a copy of version 2 (I guess no-one else
is using it yet). >>

The version 1 driver should be usable with version 2 without any changes.  I
can't remember any breaking changes to the SQL, etc. that would affect them.

--
Tim Young
Elevate Software
www.elevatesoft.com

Tim Young [Elevate Software] wrote:

>The version 1 driver should be usable with version 2 without any changes.  
>I can't remember any breaking changes to the SQL, etc. that would affect
>them.

It's the connection string that I'm having trouble getting to work, which
then requests a list of tables.

--
Jeremy Knowles

Tim Young [Elevate Software] wrote:

><< Great, thanks, presumably I just uninstall the one I have in 2007 and
>download/install the unicode one then? >>
>
>Yes.

OK, I did that, now I can't open the database as I'm getting the following
error:

ElevateDB Error #100 There is an error in the metadata for the
configuration DocsConfig (Signature, password, character set
(ANSI/Unicode), or version number mismatch)


There must be a way, but I can't see it, and there's no migrator thingy

--
Jeremy Knowles

Jeremy


The only way I can think of is to use the ansi version of EDBManager to reverse engineer the database with data, then use the unicode version of EDBManager to create the database from the script - I haven't tried it but hopefully it would work.

Roy Lambert [Team Elevate]

Jeremy,

<< OK, I did that, now I can't open the database as I'm getting the
following error: >>

That's normal - you can't interchange the ANSI with the Unicode databases.
It's either one or the other.  This is also an issue with the migrators,
which is why we don't allow for migration from ANSI to Unicode, or
vice-versa.  The actual migrator DLLs themselves are also distinctly either
ANSI or Unicode, including all of the migrator calls into the DLLs.

Roy's suggestion is the solution.  You need to reverse-engineer the tables
into an SQL script, change any ANSI collations to UNI, and then run the SQL
script against the new Unicode database.  It's a PIA, but it's the only
solution for now.

Eventually we're expecting ANSI to slowly fade away as 16-bit Unicode
becomes the standard character set in use, which is why the design was done
this way.

--
Tim Young
Elevate Software
www.elevatesoft.com

Tim Young [Elevate Software] wrote:

>
>Roy's suggestion is the solution.  You need to reverse-engineer the tables
>into an SQL script, change any ANSI collations to UNI, and then run the
>SQL script against the new Unicode database.  It's a PIA, but it's the
>only solution for now.

Righto! I guess you could get the manager to automate this in the
background, but I only need to do it this once so I don't mind (and just
to save me re-doing what I've done once).

>
>Eventually we're expecting ANSI to slowly fade away as 16-bit Unicode
>becomes the standard character set in use, which is why the design was
>done this way.

OK, well it might be nice to have some explanation like this on the
website saying something like, use the unicode one is preferred unless you
have a reason not to, since it was only at the point of downloading after
I bought it that I discovered there was a choice, but couldn't find
anything to describe the differences.

--
Jeremy Knowles

Roy Lambert wrote:

>Jeremy
>
>
>The only way I can think of is to use the ansi version of EDBManager to
>reverse engineer the database with data, then use the unicode version of
>EDBManager to create the database from the script - I haven't tried it but
>hopefully it would work.


OK, thanks, I'll try that (when I've gone back to unicode, I had to revert
back so I could do a demo of something I'd started).

--
Jeremy Knowles

Tim

>Eventually we're expecting ANSI to slowly fade away as 16-bit Unicode
>becomes the standard character set in use, which is why the design was done
>this way.

Even I do, but my definition of slowly may not include my lifetime. I wonder how much COBOL is still out there

Roy Lambert