Hello libpqxx-ML,
i'm not sure if i misunderstood something or if i did something
wrong, can you help me with this?
I wrote a little program using libpqxx-2.5.1 on amd64. I linked with
-lpqxx -lpq. It uses the tablereader-feature. When I installed the
program on a x86-machine (pentium4), my little program segfaulted. So
I talked with gdb and it seems that tablereader takes one loop longer
on x86.
Code:
vector<std::string> vtpl;
connection cn("");
work copy_n_dump(cn, "fetch_rp_los");
copy_n_dump.exec(drop_tmp_table);
copy_n_dump.exec(create_and_insert_into_tmp_table);
tablereader rd(copy_n_dump, tmp_tname);
do {
vtpl.clear();
rd >> vtpl;
t.mutate_by_iterator(vtpl.begin(), vtpl.end());
ltpls.push_back(t);
} while (rd);
rd.complete();
copy_n_dump.commit();
The segfault appeared in t.mutate_by_iterator() because i didn't check
whether vtpl.empty(). In gdb i saw an emtpy vector vtpl.
On the other machine (amd64) everything was fine without the
check.
Should it be like that?
Best Regards,
Johannes Brügmann
--
_______________________________________________
Libpqxx-general mailing list
Libpqxx-general@gborg.postgresql.org
http://gborg.postgresql.org/mailman/listinfo/libpqxx-general
