Thread

1. Re: [INTERFACES] refint doesn't work well with BDE (fwd)

   Jan Wieck <wieck@debis.com> — 2000-04-05T12:20:10Z

   > Hi Jan, refint error reporting seems to have a bug. The Windows ODBC
   > driver cannot detect that refint machine dropped an error (I think that
   > the same behaviour can be reproduced with the Unix ODBC driver, too). I
   > suspect that the 7.0 refint manager's error string has an extra \n
   > character (I found a lonely apostrophe at the beginning of a row in the
   > log). beta3 has the same problem. What's your opinion? Regards, Zoltan
   >
   > ---------- Forwarded message ----------
   > [...]
   >
   > BDE says nothing to this. BDE thinks that everything is all right, it changes
   > the table (inserts a row). After making a refresh, one can realize that no
   > changes were made, of course.
   >
   >
   > Could you please help?
   
       I'm in no way an ODBC or BDE guy - anyway.
   
       First of all, 6.5.* didn't had the FOREIGN KEY implementation
       your 7.0 message is coming from. It's completely new stuff.
   
       Second, since an ERROR is correctly reported (violating a  RI
       constraint  is an ERROR and I cannot see any extra \n there),
       the bug seems to be in the ODBC or BDE part, not  looking  at
       the  return  code  but  trying to interpret the textual error
       message.
   
   
   Jan
   
   --
   
   #======================================================================#
   # It's easier to get forgiveness for being wrong than for being right. #
   # Let's break this rule - forgive me.                                  #
   #========================================= wieck@debis.com (Jan Wieck) #
   
   
   
   

2. Re: refint doesn't work well with BDE

   Kovacs Zoltan Sandor <tip@pc10.radnoti-szeged.sulinet.hu> — 2000-04-05T15:17:46Z

   On Wed, 5 Apr 2000, Jan Wieck wrote:
   >     First of all, 6.5.* didn't had the FOREIGN KEY implementation
   >     your 7.0 message is coming from. It's completely new stuff.
   Yes, you're right, sorry. I used the contrib/spi/refint* stuff in 6.5.2.
   
   >     Second, since an ERROR is correctly reported (violating a  RI
   >     constraint  is an ERROR and I cannot see any extra \n there),
   >     the bug seems to be in the ODBC or BDE part, not  looking  at
   >     the  return  code  but  trying to interpret the textual error
   >     message.
   Checking the ODBC source, I found two places in interfaces/odbc/info.c
   with check_foreign_key and check_primary_key functions. They are the
   names of the refint check functions in the "contrib" solution in 6.5.*
   These places are queries which should be changed a bit, I think. Are these
   changes enough to make it work, Byron?
   
   Thank you again.
   
   Regards, Zoltan