
    Still to do:
        
        - Handle nested transactions.

        - Implement NLS_LANG and see brothers.

        - Implement dictionaryWithObjects:forAttributes:zone:.


Oracle7Adaptor.m:303: SEE: This implementation is insufficient. Let's have a look at Oracle Doc.
Oracle7Channel.m:466: SEE: Do we use [column typeName] or [column dataType]?
Oracle7Channel.m:475: SEE: Localisation and formating of this message.
Oracle7Channel.m:476: SEE: Should we raise an exception?
Oracle7Channel.m:483: SEE: Localisation and formating of this message.
Oracle7Channel.m:484: SEE: Should we raise an exception?
Oracle7Channel.m:525: SEE: fetchColumns is not used anywere. We should avoid creating it, the columns being retained by the buffers.
Oracle7Channel.m:718: SEE: Here, we should filter out the attributes' values
Oracle7Channel.m:719: SEE: that are not in the requested set of attributes.
Oracle7Channel.m:759: SEE: how do adaptors developed by NeXT behave respecting this point.
Oracle7Channel.m:856: SEE: Can we find a better name for generated model?
Oracle7Channel.m:863: SEE:
Oracle7Context.m:107: SEE: What can we do for nested transactions.
    
    


2000-02-22  Pascal J. Bourguignon  <pjb@imaginet.fr>

    Oracle7Adaptor v. 1.0b343

    - describeTableNames, describeModelWithTableNames:,
      readTypeForAttribute:, and readTypesForEntity corrected.

    - Contrarily  to the  previous versions, when  the select  is done
      directly      with      evaluateExpression:      instead      of
      selectAttributes:..., then fetchAttributes:withZone: will return
      a  row with anonymous  keys (@"Attribute0",  @"Attribute1", ...)
      instead     of     the      attribute     names     given     to
      fetchAttributes:withZone:,  and   like  it's  done   with  EOF-1
      adaptors written by NeXT Computer Inc.

    
2000-02-21  Pascal J. Bourguignon  <pjb@imaginet.fr>

    Oracle7Adaptor v. 1.0b318

    - Completely rewrote the fetching code. 

    - Some intense debugging.

 
2000-02-11  Pascal J. Bourguignon  <pjb@imaginet.fr>

        Cleanup for submiting to the GNUstep CVS repository.


2000-02-07  Pascal J. Bourguignon  <pjb@imaginet.fr>

        Oracle7Adaptor v. 1.0b183

    


1999-09-10  Pascal J. Bourguignon  <pjb@imaginet.fr>

        Oracle7Adaptor v. 0.4d167

    - Finished the implementation (the meta-data loading).


1999-09-02  Pascal J. Bourguignon  <pjb@imaginet.fr>

        Oracle7Adaptor v. 0.4d130

    - Corrected the SEGV that appears when deallocating  autorelease pool.
      (a copy message was missing in EOAttribute initWithName:).


1999-08-11  Pascal J. Bourguignon  <pjb@imaginet.fr>

        Oracle7Adaptor v. 0.4d107

    This is a fix release, just after having:

    - Corrected the bug I had after SELECT.
    - DELETE works;
    - INSERT works;
    - UPDATE works;
    - SELECT works;
    - hasValidConnectionDictionary implemented;

    - Remains a SEGV that appears when deallocating  autorelease pool.


1999-08-09  Pascal J. Bourguignon  <pjb@imaginet.fr>

        Oracle7Adaptor v. 0.4d93

    Everything should be implemented at the same level than the Oracle8 
    Adaptor, but there is a bug that occurs upon fetching.

    - DELETE works,
    - INSERT works,
    - UPDATE should works, but the test hang at the SELECT FOR UPDATE 
      just before.


1999-08-07  Pascal J. Bourguignon  <pjb@imaginet.fr>

        Oracle8Adaptor v. 0.6d29

    This is only a fix of the current state. 

    The adaptor has been renamed from 'Oracle' to 'Oracle8', since this
    is based on the OCI v.8 API.

    A branch will be done now to develop an 'Oracle8' adaptor, based on 
    the OCI v.7 API.

------------------------------------------------------------------------
