Re: SQL_DATA_AT_EXEC processing broken for large objects in nightly git

Hi,

I was able to reproduce it here.
Because I can't understand the current code at once, please give me some
time
to solve the problem.

regards,
Hiroshi inoue

On 2015/10/15 21:00, Vadym Krevs wrote:
> Hi,
>
> the nightly git of the psqlODBC driver appears to no longer support
> insertion into LO columns when using SQL_DATA_AT_EXEC mode.  The
> latest released driver version - psqlodbc-09.03.0400 - works.  I've
> tested this against both PostgreSQL 9.4.4 and 9.4.5 on openSUSE 13.2.
>
> An adapted variant of the large-object-test.c program that uses
> SQL_DATA_AT_EXEC mode is attached. When run as part of the regression
> test suite against psqlodbc-09.03.0400, both the
> large-object-data-at-exec-test.c test completes successfully, just
> like large-object-test.c, and produces the following output in
> test/results/large-object-data-at-exec.out:
>
> \! "./src/large-object-data-at-exec-test"
> connected
> inserting large object...
> reading it back...
> hex: 0102030405060708
> disconnecting
>
> However, when run as part of the regression test suite against nightly
> git build of psqlodbc, the large-object-data-at-exec-test.c test
> fails, unlike large-object-test.c,  and produces the following output
> in test/results/large-object-data-at-exec.out:
>
> connected
> inserting large object...
> SQLParamData failed (final call)
> HY000=unexpected NULL parameter value
>
> Regards,
> Vadym