Re[4]: [PEAR] [QF] - add rule - callback rule usingbuilt-in/registered rule inside

Tuesday, January 27, 2004, 12:50:25 PM, you wrote: >> -----Original Message----- >> From: Bertrand Mansion [mailto:bmansion@mamasam.com] >> >> <Sebastian.Baran@dialcom.com.pl> wrote : >> >> > >> > >> > Tuesday, January 27, 2004, 11:40:20 AM, you wrote: >> > >> >>> -----Original Message----- >> >>> From: Sebastian Baran [mailto:Sebastian.Baran@dialcom.com.pl] >> >>> >> >>> I want to create custom callback rule. But part of it's >> functionality >> >>> is going to be the same as one of built-in rules. My >> question is - is >> >>> it possible to call built-in rule from my callback function? Or >> >>> do I have to rewrite built-in rule functionality by hand in my >> >>> function? >> > >> > MC> Depending on exactly what your rule is going to do, >> > MC> could you not add the rules in the right order, so that >> the built-in one >> > is >> > MC> called before your custom callback one? >> > >> > MC> - Mike >> > >> > >> > Well, I didn't think about it this way. It's good idea - I'll try to >> > put two rules instead of one complex rule. This will be >> sufficient for >> > most cases. >> > >> > But if I would like to have just one rule to overwrite >> error messages >> > generated by two rules and display just one error message? For >> > example: >> > - one rule - built-in 'email' to initially check if email is ok >> > - second rule - callback to check something more complex >> (for example >> > if email is in DB or something). >> > >> > Is it good way to do it? Can I put both rules into one callback rule >> > calling buil-in rule first? >> >> I don't see any reason why you would want to do that ? MC> If the email address "ok" check (via regex) is done on the client side MC> before submitting the data to the server this should be fine. Because the MC> second check (to see if its in the dbase) would not happen unless phase 1 MC> was passed. MC> However if all validation is server side only - ie regex on the server, MC> followed by a DB lookup, then two error messages would be returned - in MC> effect telling the user the same thing twice! MC> - The email address is not valid MC> - The email address is not in the database MC> Doesn’t look too great as one infers the other. Computers are supposed to MC> be clever ;-) Well, it seems you're right about it :-) Thanks for helping me to see it from other side :-) -- Best regards, Sebastian