Tim Bunce, author of the Perl DBI, shares your reasoning. In the spirit of perl (there's more than one way to do things), he offered both a virtualized error message system (ODBC-like) and also the proprietary DB's error numbers and messages. Of course, people used the proprietary error numbers and messages almost exclusively, for the reasons you give. cheers, monty > -----Original Message----- > From: John Lim [mailto:[EMAIL PROTECTED]] > Sent: Wednesday, May 16, 2001 10:17 AM > To: [EMAIL PROTECTED] > Subject: Re: [PHP-WIN] the /php/pear/DB abstraction > > > Hello Monty, > > As an author of a database class library > (http://php.weblogs.com/adodb) I > can understand the need for good error messages. > > I decided in adodb to avoid using virtualised error messages > because of this > issue. It's more important to give good error messages than > attempting to > map > to a virtualised error message system. Why? You normally just > rollback on > any error anyway, and the real error messages allow you to pinpoint > the bug precisely. > > You might like to give it a test run. > > Regards, John > -- PHP Windows Mailing List (http://www.php.net/) To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] To contact the list administrators, e-mail: [EMAIL PROTECTED]
