Re: RE : RE : problems with back-sql

["Pierangelo Masarati" <ando@sys-net.it>]

>>
>>> - ITS#2642 :  tested, OK, though another problem occurs later in
>>> backsql_count_children(), in the call to backsql_bindRowAsStrings()
> (I
>>> have to investigate this problem more).
>>
>>I'll also look at it.
>>
>>p.
>
> While on the problem, I gdb'd it a little.
> Here is what I saw: with oracle, the result from the SELECT COUNT(...)
> request is typed SQL_NUMERIC (in backsql_bindRowAsStrings(), col_type==
> 2). So when there is no child, rows.cols[0] is set to "0.0000000" and
> the strtol() call in backsql_count_children() sets the end pointer to
> '.' and not to '\0'. The following test fails and the result is set to
> LDAP_OTHER...

I guess my test was a bit overkill (although I really expect
count to return an integer ...); a quick fix could be
to suppress the test on the next[0] value returned by strto[u]l ...

> I don't know why the COUNT colum is typed SQL_NUMERIC in oracle and
> SQL_INTEGER in MySQL...

p.

-- 
Pierangelo Masarati
mailto:pierangelo.masarati@sys-net.it