You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@commons.apache.org by James Turner <tu...@blackbear.com> on 2002/09/19 19:40:49 UTC
VOTE: Behavior of Validator
As currently written, the Validator has what I consider a quirk.
Suppose you have two fields, username and password. Username has
"depends=required" and password has "depends=required.notgod" (where notgod
is a test that makes sure that the user didn't choose god as a
password). The following behavior occurs:
username=blank, password=blank: two errors generated on "required"
username=blank, password=god: one error generated on "required"
username=george, password=god: one error genereated on "notgod"
This is because the Validator won't look at notgod until *all* fields pass
the "required" test.
I think this is a broken behavior. It leads to web forms where the user
thinks that they've filled in all the fields correctly, but then get new
error messages they've never seen before. I'd like to correct this before
Validator freezes for a release, but I want to make sure no one really
really thinks that the current behavior is somehow the right one. So
please vote:
+1 = Change Validator so that this doesn't occur
0 = I don't give a hoot
-1 = I really like the way it works now (please give a reason)
I'm sending this both to Commons and Struts because both communities are
impacted by the change.
James
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
Re: VOTE: Behavior of Validator
Posted by "V. Cekvenich" <vi...@users.sourceforge.net>.
+1 and thanks!!!!
V.
James Turner wrote:
> As currently written, the Validator has what I consider a quirk.
>
> Suppose you have two fields, username and password. Username has
> "depends=required" and password has "depends=required.notgod" (where
> notgod is a test that makes sure that the user didn't choose god as a
> password). The following behavior occurs:
>
> username=blank, password=blank: two errors generated on "required"
> username=blank, password=god: one error generated on "required"
> username=george, password=god: one error genereated on "notgod"
>
> This is because the Validator won't look at notgod until *all* fields
> pass the "required" test.
>
> I think this is a broken behavior. It leads to web forms where the user
> thinks that they've filled in all the fields correctly, but then get new
> error messages they've never seen before. I'd like to correct this
> before Validator freezes for a release, but I want to make sure no one
> really really thinks that the current behavior is somehow the right
> one. So please vote:
>
> +1 = Change Validator so that this doesn't occur
> 0 = I don't give a hoot
> -1 = I really like the way it works now (please give a reason)
>
> I'm sending this both to Commons and Struts because both communities are
> impacted by the change.
>
> James
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
Re: VOTE: Behavior of Validator
Posted by Dave Derry <dd...@acm.org>.
----- Original Message -----
From: "James Turner" <tu...@blackbear.com>
>
> Precisely. Validation for a field should stop on the first error *for
that
> field*
>
> James
>
+1
I agree whole-heartedly with this. The way the Validator currently works is
to iterate throught the validation rules, and for each rule iterate through
the fields to locate those that depend on that rule. This seems wasteful to
me. I think a better way would be to iterate through the list of fields, and
for each field execute the validation methods for that field. This should
accomplish what you state above.
Dave D
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
RE: VOTE: Behavior of Validator
Posted by James Turner <tu...@blackbear.com>.
At 02:02 PM 9/19/2002, Dominique Plante wrote:
>+1
>
>Good example, and I agree with your rationale.
>
>Under the new behavior following the example you gave for the second
>case:
>username=blank, password=god: one error generated on "required"
>
>Two errors should be generated:
>username is required
>password cannot be god
>
>Is the new behavior that you want to have happen?
Precisely. Validation for a field should stop on the first error *for that
field*
James
>Dominique
>
>
>-----Original Message-----
>From: James Turner [mailto:turner@blackbear.com]
>Sent: Thursday, September 19, 2002 10:41 AM
>To: commons-dev@jakarta.apache.org
>Cc: struts-dev@jakarta.apache.org
>Subject: VOTE: Behavior of Validator
>
>As currently written, the Validator has what I consider a quirk.
>
>Suppose you have two fields, username and password. Username has
>"depends=required" and password has "depends=required.notgod" (where
>notgod
>is a test that makes sure that the user didn't choose god as a
>password). The following behavior occurs:
>
>username=blank, password=blank: two errors generated on "required"
>username=blank, password=god: one error generated on "required"
>username=george, password=god: one error genereated on "notgod"
>
>This is because the Validator won't look at notgod until *all* fields
>pass
>the "required" test.
>
>I think this is a broken behavior. It leads to web forms where the user
>
>thinks that they've filled in all the fields correctly, but then get new
>
>error messages they've never seen before. I'd like to correct this
>before
>Validator freezes for a release, but I want to make sure no one really
>really thinks that the current behavior is somehow the right one. So
>please vote:
>
>+1 = Change Validator so that this doesn't occur
>0 = I don't give a hoot
>-1 = I really like the way it works now (please give a reason)
>
>I'm sending this both to Commons and Struts because both communities are
>
>impacted by the change.
>
>James
>
>
>
>--
>To unsubscribe, e-mail:
><ma...@jakarta.apache.org>
>For additional commands, e-mail:
><ma...@jakarta.apache.org>
>
>
>--
>To unsubscribe, e-mail: <ma...@jakarta.apache.org>
>For additional commands, e-mail: <ma...@jakarta.apache.org>
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
Re: VOTE: Behavior of Validator
Posted by "V. Cekvenich" <vi...@users.sourceforge.net>.
+1 and thanks!!!!
V.
James Turner wrote:
> As currently written, the Validator has what I consider a quirk.
>
> Suppose you have two fields, username and password. Username has
> "depends=required" and password has "depends=required.notgod" (where
> notgod is a test that makes sure that the user didn't choose god as a
> password). The following behavior occurs:
>
> username=blank, password=blank: two errors generated on "required"
> username=blank, password=god: one error generated on "required"
> username=george, password=god: one error genereated on "notgod"
>
> This is because the Validator won't look at notgod until *all* fields
> pass the "required" test.
>
> I think this is a broken behavior. It leads to web forms where the user
> thinks that they've filled in all the fields correctly, but then get new
> error messages they've never seen before. I'd like to correct this
> before Validator freezes for a release, but I want to make sure no one
> really really thinks that the current behavior is somehow the right
> one. So please vote:
>
> +1 = Change Validator so that this doesn't occur
> 0 = I don't give a hoot
> -1 = I really like the way it works now (please give a reason)
>
> I'm sending this both to Commons and Struts because both communities are
> impacted by the change.
>
> James
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
RE: VOTE: Behavior of Validator
Posted by Dominique Plante <dp...@pacbell.net>.
+1
Good example, and I agree with your rationale.
Under the new behavior following the example you gave for the second
case:
username=blank, password=god: one error generated on "required"
Two errors should be generated:
username is required
password cannot be god
Is the new behavior that you want to have happen?
Dominique
-----Original Message-----
From: James Turner [mailto:turner@blackbear.com]
Sent: Thursday, September 19, 2002 10:41 AM
To: commons-dev@jakarta.apache.org
Cc: struts-dev@jakarta.apache.org
Subject: VOTE: Behavior of Validator
As currently written, the Validator has what I consider a quirk.
Suppose you have two fields, username and password. Username has
"depends=required" and password has "depends=required.notgod" (where
notgod
is a test that makes sure that the user didn't choose god as a
password). The following behavior occurs:
username=blank, password=blank: two errors generated on "required"
username=blank, password=god: one error generated on "required"
username=george, password=god: one error genereated on "notgod"
This is because the Validator won't look at notgod until *all* fields
pass
the "required" test.
I think this is a broken behavior. It leads to web forms where the user
thinks that they've filled in all the fields correctly, but then get new
error messages they've never seen before. I'd like to correct this
before
Validator freezes for a release, but I want to make sure no one really
really thinks that the current behavior is somehow the right one. So
please vote:
+1 = Change Validator so that this doesn't occur
0 = I don't give a hoot
-1 = I really like the way it works now (please give a reason)
I'm sending this both to Commons and Struts because both communities are
impacted by the change.
James
--
To unsubscribe, e-mail:
<ma...@jakarta.apache.org>
For additional commands, e-mail:
<ma...@jakarta.apache.org>
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
RE: VOTE: Behavior of Validator
Posted by phpsurf <ph...@ifrance.com>.
+1 definitely
I definitely agree with both:
- The validation process should first iterate through all fields and then,
for each field, iterate through its scpecific validation rules.
- The validation process can stop (for each field) at the first unvalidated
rule, but it should not stop at the first unvalidated field ...
> -----Original Message-----
> From: James Turner [mailto:turner@blackbear.com]
> Sent: jeudi 19 septembre 2002 19:41
> To: commons-dev@jakarta.apache.org
> Cc: struts-dev@jakarta.apache.org
> Subject: VOTE: Behavior of Validator
>
>
> As currently written, the Validator has what I consider a quirk.
>
> Suppose you have two fields, username and password. Username has
> "depends=required" and password has "depends=required.notgod"
> (where notgod
> is a test that makes sure that the user didn't choose god as a
> password). The following behavior occurs:
>
> username=blank, password=blank: two errors generated on "required"
> username=blank, password=god: one error generated on "required"
> username=george, password=god: one error genereated on "notgod"
>
> This is because the Validator won't look at notgod until *all*
> fields pass
> the "required" test.
>
> I think this is a broken behavior. It leads to web forms where the user
> thinks that they've filled in all the fields correctly, but then get new
> error messages they've never seen before. I'd like to correct
> this before
> Validator freezes for a release, but I want to make sure no one really
> really thinks that the current behavior is somehow the right one. So
> please vote:
>
> +1 = Change Validator so that this doesn't occur
> 0 = I don't give a hoot
> -1 = I really like the way it works now (please give a reason)
>
> I'm sending this both to Commons and Struts because both communities are
> impacted by the change.
>
> James
>
>
>
> --
> To unsubscribe, e-mail:
> <ma...@jakarta.apache.org>
> For additional commands, e-mail:
> <ma...@jakarta.apache.org>
>
________________________________________________________________
Etudiant: Wanadoo t'offre le Pack eXtense Haut D�bit soit 150,92 euros d'�conomies !
Et pour 1 euro de plus, re�ois le CD-ROM du jeu Dark Age of Camelot
+ 1 mois de jeu en r�seau offert !
Clique ici : http://www.ifrance.com/_reloc/mail.etudiant
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>
Re: VOTE: Behavior of Validator
Posted by di...@multitask.com.au.
James Turner <tu...@blackbear.com> wrote on 20/09/2002 03:40:49 AM:
> As currently written, the Validator has what I consider a quirk.
>
> Suppose you have two fields, username and password. Username has
> "depends=required" and password has "depends=required.notgod" (where
notgod
> is a test that makes sure that the user didn't choose god as a
> password). The following behavior occurs:
>
> username=blank, password=blank: two errors generated on "required"
> username=blank, password=god: one error generated on "required"
> username=george, password=god: one error genereated on "notgod"
>
> This is because the Validator won't look at notgod until *all* fields
pass
> the "required" test.
>
> I think this is a broken behavior. It leads to web forms where the user
> thinks that they've filled in all the fields correctly, but then get new
> error messages they've never seen before. I'd like to correct this
before
> Validator freezes for a release, but I want to make sure no one really
> really thinks that the current behavior is somehow the right one. So
> please vote:
>
> +1 = Change Validator so that this doesn't occur
> 0 = I don't give a hoot
> -1 = I really like the way it works now (please give a reason)
>
> I'm sending this both to Commons and Struts because both communities are
> impacted by the change.
+1
--
dIon Gillard, Multitask Consulting
Work: http://www.multitask.com.au
Developers: http://adslgateway.multitask.com.au/developers
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>