For your reference:
sub dump_results { # also aliased as a method in DBD::_::st
my ($sth, $maxlen, $lsep, $fsep, $fh) = @_;
return 0 unless $sth;
$maxlen ||= 35;
$lsep ||= "\n";
$fh ||= \*STDOUT;
my $rows = 0;
my $ref;
while($ref = $sth->fetch) {
print $fh $lsep if $rows++ and $lsep;
my $str = neat_list($ref,$maxlen,$fsep);
print $fh $str; # done on two lines to avoid 5.003 errors
}
print $fh "\n$rows rows".($DBI::err ? " ($DBI::err: $DBI::errstr)" :
"")."\n";
$rows;
}
I would like to add another parameter that would suppress the rowcount print
as long as there are no errors. In theory:
sub dump_results { # also aliased as a method in DBD::_::st
my ($sth, $maxlen, $lsep, $fsep, $fh, $suppress_rowcount) = @_;
...
print $fh "\n$rows rows".($DBI::err ? " ($DBI::err: $DBI::errstr)" :
"")."\n"
if(not defined($suppress_rowcount) ## default behaviour
or not $suppress_rowcount ## if 1, it may suppress
or $DBI::err ## print if there's an error,
regardless
);
$rows;
}
In practice, I think that I'd also like to add also add $suppress_errors.
Seems to me that what we should really do is:
sub dump_results { # also aliased as a method in DBD::_::st
my ($sth, $maxlen, $lsep, $fsep, $fh, $hr_attribs) = @_;
## hr_ is "hungarian" for hash ref. Not neccessary, but for this
example, I'll use it for clarity
## Pedantic, but I ABHOR 'use-of-uninitialized' warnings...
my $suppress_count = (defined($hr_attribs)
and defined($hr_attribs->{'suppress_count'}
and $hr_attribs->{'suppress_count'});
my $suppress_errors = (defined($hr_attribs)
and defined($hr_attribs->{'suppress_errors'}
and $hr_attribs->{'suppress_errors'});
...
print $fh "\n$rows rows" if (not $suppress_count);
if($DBI::err and not $suppress_errors) {
print $fh ($suppress_count ? "\n" : " ");
print $fh "($DBI::err: $DBI::errstr)";
}
print $fh "\n";
$rows;
}
This arrangement will allow us to (for example) add on formating specs to
neat_list in the same manner and pass them down.
Can I get a consensus and critique before I submit a patch?
--
Matthew O. Persico
