ColdFusion 10 - No MS Access with Unicode driver

Please vote on this bug https://bugbase.adobe.com/index.cfm?event=bug&id=3691289

It stops us from migrating from CF 9 to 10 as we of course don't have 32 servers anymore.

Hello IcePerson13,

  this is possible. However, it is only possible within the 32bit Version of ColdFusion on a 64bit OS and - amazingly - not in the standard ColdFusion 10 32bit setup routine on a 64bit OS. That there is no support for 64bit CF10 on a 64bit OS is likely due to the ODBC limitations.

The method is described at:

http://cfassociates.samuraiz.co.jp/index.cfm/faq/aboutcf/cf10-nosupport-msaccess-unicode/

(open this url in Google Chrome, which will then offer you automatically a reasonable translation).

The trick is, to install CF10 32bit on a 32bit OS, be it WinXP or whatever. Then pickup the izmjniado.dll from that machine at [cf10_root]/cfusion/lib. This is the dll which offers the connection "Microsoft Access with Unicode".

Then install CF10 32bit on your 64bit OS.

Then copy the izmjniado.dll to [cf10_root] / cfusion / lib on the 64bit machine you actually have installed the 32bit CF10 on (stop CF service before, of course). After restarting CF, the datasource option "Microsoft Access with Unicode" is already there, but will throw an error when you want to bind in a connection to a MSAccess database on your machine.

Then correct this error cause via editing neo-datasource.xml, by cleaning it up (first stop CF server and make a backup of the original file). Basically get rid of the "<char code="09">" strings in this xml file. No line break in the middle, one space between "Data" and "Source".

so before it looks like:

• <var name='url'> <string> jdbc: izmado: Provider = Microsoft.Jet.OLEDB.4.0; Data <char code='0a'/> <char code='09'/> <char code = '09 '/> <char code='09'/> <char code='09'/> <char code='09'/> <char code='09'/> <char code='09'/> <char code = '09 '/> Source = C :/ db / uriage.mdb; IzmJdbcEsc = yes; IzmReleaseOnClose = no <char code='0a'/> <char code='09'/> <char code = '09' /> <char code='09'/> <char code='09'/> <char code='09'/> <char code='09'/> <char code='09'/> </ string > </ var>

afterwards it should look like

• <var name='url'> <string> jdbc: izmado: Provider = Microsoft.Jet.OLEDB.4.0; Data Source = C :/ db / uriage.mdb; IzmJdbcEsc = yes; IzmReleaseOnClose = no </ string> </ var>

Then restart CF10 and there you are. We have it running wonderfully against Access with Unicode on a Win2012R, Apache 2.

Why Adobe Systems ignores this way a *large* part of its customers who need to use - be it only for legacy reasons - MSAccess with Unicode, remains a complete mistery for anyone thinking half way commercially. It is even more hallucinating because Adobe Systems itself declares in its specs that CF10 works with MSAccess containing Unicode strings (see: the CF10 specification document from Adobe itself, where Adobe solemly declares

"Connecting to Microsoft Access with Unicode" at http://help.adobe.com/en_US/ColdFusion/10.0/Admin/WSc3ff6d0ea77859461172e0811cbf364104-7fe6.html)/WSc3ff6d0ea77859461172e0811cbf364104-7fe6.html). It may well be that this Unicode feature is not very relevant for the American market (the standard MSAccess driver does connect to MSAccess but does not deal with its Unicode strings), but it definitely is for the rest of the world.

Dealing like this with extremely popular Unicode enabled products feels like having been thrown back 10 years back in the ColdFusion development.

So, yes, this works. But Adobe should fix this immediately in its install routines.