FreeTDS Development Group

["James K. Lowden" <jklowden AT schemamania.org>]

"Craig A. Berry" <craigberry AT mac.com> wrote:
> I would guess the normal method of exit is up above where it says:
> 
>                         free_bread_crumb();
>                         return (last_read == 48) ? 0 : 1;
> 
> Why it doesn't hit that now I don't know.

I think the test is broken.  That shouldn't surprise me, because it passed
when buffering was broken.  But let me push it around with you, if I may.

The test hasn't changed much since v 1.1, 3 1/2 years ago:

rows_to_add = 50;
...
        for (i = 1; i < rows_to_add; i++) {
...
                if (i % 25 == 0) {
                        dbclrbuf(dbproc, 25);
                }
                if (REG_ROW != dbnextrow(dbproc)) {
...                     [... "normal exit" ...]
        }
...
        return 1;

It will retrieve only 49 of the 50 rows selected.  dbnextrow() will always
return REG_ROW, because it never exhausts the resultset.  Am I missing
something?  

I'm tempted to remove this test, because t0002 does a more thorough job. 
But for now I'm going to commit what I've got.  

Regards, 

--jkl