Re: [mdb-dev] Bug in mdb-export producing NULL values

[Yasir A. <li...@en...>]

Hello,

Have any of the MDB Tools developers had a chance to look at my patch 
(below)? I sent this email in a few weeks ago.

$ cvs diff data.c
Index: data.c
===================================================================
RCS file: /cvsroot/mdbtools/mdbtools/src/libmdb/data.c,v
retrieving revision 1.102
diff -r1.102 data.c
173a174,176
 >       if (col->len_ptr) {
 >               *col->len_ptr = 1;
 >       }


Thanks,
Yasir

> Many thanks to the team for developing this code - I've found it very 
> useful.
>
> I'm using the latest CVS version on Linux
>
> I think I found a bug in data.c which causes mdb-export -I to 
> sometimes print NULL for boolean columns (even when the value is not 
> null).
>
> Here's the CVS diff:
>
> $ cvs diff data.c
> Index: data.c
> ===================================================================
> RCS file: /cvsroot/mdbtools/mdbtools/src/libmdb/data.c,v
> retrieving revision 1.102
> diff -r1.102 data.c
> 173a174,176
> >       if (col->len_ptr) {
> >               *col->len_ptr = 1;
> >       }
>
> I hardly know the code at all, so my fix could be completely wrong, 
> but it looks as though we're not updating *col->len_ptr for for 
> boolean fields and this causes the -I switch of mdb-export (which 
> relies on col->len_ptr) to think its value is 0. In fact I'm not sure 
> of the value is ever set to 0 to begin with so whether it prints NULL 
> or not depends on what's in the block of memory when bound_lens is 
> allocated (see mdb-export.c).
>
> Just in case this isn't clear, the change is in data.c, function 
> mdb_xfer_bound_bool(), with the new code inserted at line 174.
>
> Thanks,
> Yasir