in reply to Re: How to subclass DBI and still have the errors reported in user code
in thread How to subclass DBI and still have the errors reported in user code
Thanks for idea, it is the solution, but:
With catch and croak you have to take care of context (AFAIK), something like this:
sub selectrow_array { my $this = shift(); my ( @ret, $ret ); eval { if (wantarray) { # list @ret = $this->SUPER::selectrow_array(@_); } elsif ( defined(wantarray) ) { # scalar $ret = $this->SUPER::selectrow_array(@_); } else { # void - nonsense $this->SUPER::selectrow_array(@_); } }; croak $@ if $@; return if !defined wantarray; return wantarray ? @ret : $ret; }
The original error line is still being reported. Of course you can strip it with RE. But with $dbh->{'PrintError'} set the line still appear.
DBD::mysql::db selectrow_array failed: Table 'devel.no_such_table' doe +sn't exist at simple.pl line 17. DBD::mysql::db selectrow_array failed: Table 'devel.no_such_table' doe +sn't exist at simple.pl line 17. at simple.pl line 43
DBI actually "croaks" for itself. The error is reported on the boundary of user and DBI code. What I want is to tell DBI: this code is safe, don't report errors from here, go up in stack. I dont' know whether DBI has some rules for telling whether a call shouldn't generate errors as Carp have (inheritance is one the rules).
Maybe that the elegant, DBI specific solution, doesn't exist.
|
|---|
| Replies are listed 'Best First'. | |
|---|---|
|
Re^3: How to subclass DBI and still have the errors reported in user code
by moritz (Cardinal) on Sep 11, 2007 at 13:47 UTC |