first patch to odbc

Brian Bruns camber at ais.org
Wed Apr 10 21:30:08 EDT 2002



Patch is commited.  BTW, I believe you have a bug in the get_client_type 
function.  The inner case for SYBFLTN is a switch on column_type and 
should be column_size.

I fixed it while making the other change (not passing column struct) I 
mentioned in the other email.

Brian

On Tue, 9 Apr 2002, ZIGLIO Frediano wrote:

> > 
> > Thank you for your patch.  I reviewed it carefully.
> > 
> > I'm one of the CVS committers, but I'm not understanding the patch so
> > I'll leave it to our leader Brian to apply.  However, I thought you
> > might be interested in why I don't understand it, and why I 
> > decided not
> > to apply it.  Some of my reasons are that I lack experience with the
> > code, but there are other reasons.
> > 
> > - In SQLMoreResults() you check for HSTMT twice at the start of the
> > function.  I think this is wrong.
> > 
> 
> Opps... I made an error merge some fix (first time I patched 0.53, 
> not current)
> 
> > - Many functions you changed so that they return failure instead of
> > success.  You didn't say why this change was done.  Do you think this
> > might break existing code in other projects?
> > 
> > - You didn't include a patch to ChangeLog, NEWS or to any 
> > documentation
> > files.  Whoever applies the patch for you would have to go through all
> > other files to see if they have to be changed.
> 
> Ok, written in ChangeLog. Reported for ML convenience:
> * src/odbc/odbc.c:
> - changed return of unimplemented function from SQL_SUCCESS to
> SQL_ERROR. User program assume that result are successfully set on
> SQL_SUCCESS so user program can read garbage data or do other strange
> assuming. Not changed result for function like SQLDisconnect that
> work in the same way
> - implemented SQLMoreResults and SQLNumParams
> - return SQL_ERROR on memory allocation failure
> - check all statement entering in functions
> - removed some compiler warning due to conversion
> - improved SQLFreeStmt (support all type of free) and fix (not unbind
> on SQL_CLOSE)
> - added some type handling (like SYBINTN) converting from tds to odbc
> - fixed some strange indentation
> 
> > 
> >- The use of // as a comment marker is not K&R.
> >
> 
> I use mainly C++, sometime ANSI C, what are main differency from K&R and
> ANSI ?
> 
> >I suggest you have a read of this chapter:
> >
> >http://www.linuxdoc.org/HOWTO/Software-Release-Practice-HOWTO/patching.html
> >
> 
> Readed
> 
> I updated my patch with your suggestion.
> 
> bye
>   freddy77
> 
> 




More information about the FreeTDS mailing list