[email protected] (Rudy Lippan) wrote:
>>> an implementation problem in DBD::Pg, but I am not convinced. What does
>>> a DBI->trace(9) show?
>>> [...]
>> | DBI->trace (9);
>> | my $DB = DBI->connect ('dbi:Pg:dbname=tim', undef, undef, { pg_enable_utf8
>> => 1, PrintError => 0 }) or die (DBI->errstr ());
>> | my $TestOutput = $DB->selectcol_arrayref ('SELECT
>> LENGTH(unnest(?::TEXT[]));', {}, ['abc', 'def', 'g\hi']) or die ($DB->errstr
>> ());
>> | $DB->disconnect () or die ($DB->errstr ());
>> | print Dumper $TestOutput;
>> gives about 300 lines of output:
> [...]
>> | Begin dbd_bind_ph (ph_name: 1)
>> | Begin pg_stringify_array
>> | End pg_stringify_array (string: '{"abc","def","g\\\\hi"}')
> My first impresison is that pg_stringify_array is giving you your extra
> backslash. I don't have time to dig much into it right now. Maybe
> someone who knows that bit of code could tell you what the problem is....
>> | Bind (1) (type=0)
>> | End dbd_bind_ph (special)
>> | Begin dbd_st_execute
>> | PQexecPrepared item #0
>> | -> Value: ({"abc","def","g\\\\hi"})
I think Noah's patch (cf.
<URI:http://rt.cpan.org/Public/Bug/Display.html?id=58552>)
addresses the problem, but I haven't tested it yet. Maybe
some developer could take a deeper look into it?
Tim
