[jira] Created: (MYFACES-2256) oamSetHiddenInput should not ignore nodeName=='input' (lower case characters)

[My Faces - Dev mailing list]

oamSetHiddenInput should not ignore nodeName=='input' (lower case characters)
-----------------------------------------------------------------------------

                 Key: MYFACES-2256
                 URL: https://issues.apache.org/jira/browse/MYFACES-2256
             Project: MyFaces Core
          Issue Type: Bug
    Affects Versions: 1.2.6
         Environment: Windows XP Professional, Firefox 3.0.11, MyFaces 1.2.6 Tomahawk-12 1.1.8 (and RichFaces 3.3.1 GA), Tomcat 6.
            Reporter: Lutz Ulruch


In

function oamSetHiddenInput(formname, name, value)

the test

if(typeof form.elements[name]!='undefined' && form.elements[name].nodeName=='INPUT')

does not cover <input> elements where nodeName is 'input' (that is: lower case characters instead of upper case 'INPUT').
This results in the following error:
If the form already contains an input element with the specified name, but nodeName=='input', the test fails and a new input element with the same name is added to the form.
If the form is submitted, the HTTP request will contain 2 values for the parameter referenced by 'name':
one is the value of the unchanged <INPUT> field, the other is the value of the newly created <input> field.
Most UIComponents do not check if there is more than one value for a parameter.
Probably they will just interpret the first value - which might be the 'wrong' value of the <INPUT> field instead of the 'valid' value of the <input> field.

How I achieved this situation:
Even though MyFaces typically renders input fields with lower case characters (XHTML), when a page is rendered by Firefox, the nodeName of input elements seems to be the upper case variant. I wonder why.
Now, we use RichFaces' a4j:poll in order to perform periodical partial page updates.
The updated part of the page may contain locically equivalent input fields, but in lower case.
The JavaScript behind a4j:poll uses the W3C DOM method replaceChild() in order to update page fragments.
This method replaces the original <INPUT> fields by <input> fields.
Of course, it doesn't matter how <input> elements become part of a page.
In oamSetHiddenInput() <input> should be treated like <INPUT>.

--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

[My Faces - Dev mailing list]

     [ https://issues.apache.org/jira/browse/MYFACES-2256?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Lutz Ulruch updated MYFACES-2256:
---------------------------------

    Status: Patch Available  (was: Open)

> oamSetHiddenInput should not ignore nodeName=='input' (lower case characters)
> -----------------------------------------------------------------------------
>
>                 Key: MYFACES-2256
>                 URL: https://issues.apache.org/jira/browse/MYFACES-2256
>             Project: MyFaces Core
>          Issue Type: Bug
>    Affects Versions: 1.2.6
>         Environment: Windows XP Professional, Firefox 3.0.11, MyFaces 1.2.6 Tomahawk-12 1.1.8 (and RichFaces 3.3.1 GA), Tomcat 6.
>            Reporter: Lutz Ulruch
>
> In
> function oamSetHiddenInput(formname, name, value)
> the test
> if(typeof form.elements[name]!='undefined' && form.elements[name].nodeName=='INPUT')
> does not cover <input> elements where nodeName is 'input' (that is: lower case characters instead of upper case 'INPUT').
> This results in the following error:
> If the form already contains an input element with the specified name, but nodeName=='input', the test fails and a new input element with the same name is added to the form.
> If the form is submitted, the HTTP request will contain 2 values for the parameter referenced by 'name':
> one is the value of the unchanged <INPUT> field, the other is the value of the newly created <input> field.
> Most UIComponents do not check if there is more than one value for a parameter.
> Probably they will just interpret the first value - which might be the 'wrong' value of the <INPUT> field instead of the 'valid' value of the <input> field.
> How I achieved this situation:
> Even though MyFaces typically renders input fields with lower case characters (XHTML), when a page is rendered by Firefox, the nodeName of input elements seems to be the upper case variant. I wonder why.
> Now, we use RichFaces' a4j:poll in order to perform periodical partial page updates.
> The updated part of the page may contain locically equivalent input fields, but in lower case.
> The JavaScript behind a4j:poll uses the W3C DOM method replaceChild() in order to update page fragments.
> This method replaces the original <INPUT> fields by <input> fields.
> Of course, it doesn't matter how <input> elements become part of a page.
> In oamSetHiddenInput() <input> should be treated like <INPUT>.

--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

[My Faces - Dev mailing list]

    [ https://issues.apache.org/jira/browse/MYFACES-2256?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12721709#action_12721709 ]

Lutz Ulruch commented on MYFACES-2256:
--------------------------------------

The patch for this issue seems to be quite simple:

In MyFaces
org.apache.myfaces.shared_impl.renderkit.html.HtmlRendererUtils.prepareScript(ScriptContext context, boolean autoScroll)

and Tomahawk

org.apache.myfaces.shared_tomahawk.renderkit.html.HtmlRendererUtils.prepareScript(ScriptContext context, boolean autoScroll)

replace:

context.append("if(typeof form.elements[name]!='undefined' && form.elements[name].nodeName=='INPUT')");

with:

context.append("if(typeof form.elements[name]!='undefined' && (form.elements[name].nodeName=='INPUT' || form.elements[name].nodeName=='input'))");

> oamSetHiddenInput should not ignore nodeName=='input' (lower case characters)
> -----------------------------------------------------------------------------
>
>                 Key: MYFACES-2256
>                 URL: https://issues.apache.org/jira/browse/MYFACES-2256
>             Project: MyFaces Core
>          Issue Type: Bug
>    Affects Versions: 1.2.6
>         Environment: Windows XP Professional, Firefox 3.0.11, MyFaces 1.2.6 Tomahawk-12 1.1.8 (and RichFaces 3.3.1 GA), Tomcat 6.
>            Reporter: Lutz Ulruch
>
> In
> function oamSetHiddenInput(formname, name, value)
> the test
> if(typeof form.elements[name]!='undefined' && form.elements[name].nodeName=='INPUT')
> does not cover <input> elements where nodeName is 'input' (that is: lower case characters instead of upper case 'INPUT').
> This results in the following error:
> If the form already contains an input element with the specified name, but nodeName=='input', the test fails and a new input element with the same name is added to the form.
> If the form is submitted, the HTTP request will contain 2 values for the parameter referenced by 'name':
> one is the value of the unchanged <INPUT> field, the other is the value of the newly created <input> field.
> Most UIComponents do not check if there is more than one value for a parameter.
> Probably they will just interpret the first value - which might be the 'wrong' value of the <INPUT> field instead of the 'valid' value of the <input> field.
> How I achieved this situation:
> Even though MyFaces typically renders input fields with lower case characters (XHTML), when a page is rendered by Firefox, the nodeName of input elements seems to be the upper case variant. I wonder why.
> Now, we use RichFaces' a4j:poll in order to perform periodical partial page updates.
> The updated part of the page may contain locically equivalent input fields, but in lower case.
> The JavaScript behind a4j:poll uses the W3C DOM method replaceChild() in order to update page fragments.
> This method replaces the original <INPUT> fields by <input> fields.
> Of course, it doesn't matter how <input> elements become part of a page.
> In oamSetHiddenInput() <input> should be treated like <INPUT>.

--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

[My Faces - Dev mailing list]

     [ https://issues.apache.org/jira/browse/MYFACES-2256?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Lutz Ulruch updated MYFACES-2256:
---------------------------------

    Status: Open  (was: Patch Available)

> oamSetHiddenInput should not ignore nodeName=='input' (lower case characters)
> -----------------------------------------------------------------------------
>
>                 Key: MYFACES-2256
>                 URL: https://issues.apache.org/jira/browse/MYFACES-2256
>             Project: MyFaces Core
>          Issue Type: Bug
>    Affects Versions: 1.2.6
>         Environment: Windows XP Professional, Firefox 3.0.11, MyFaces 1.2.6 Tomahawk-12 1.1.8 (and RichFaces 3.3.1 GA), Tomcat 6.
>            Reporter: Lutz Ulruch
>
> In
> function oamSetHiddenInput(formname, name, value)
> the test
> if(typeof form.elements[name]!='undefined' && form.elements[name].nodeName=='INPUT')
> does not cover <input> elements where nodeName is 'input' (that is: lower case characters instead of upper case 'INPUT').
> This results in the following error:
> If the form already contains an input element with the specified name, but nodeName=='input', the test fails and a new input element with the same name is added to the form.
> If the form is submitted, the HTTP request will contain 2 values for the parameter referenced by 'name':
> one is the value of the unchanged <INPUT> field, the other is the value of the newly created <input> field.
> Most UIComponents do not check if there is more than one value for a parameter.
> Probably they will just interpret the first value - which might be the 'wrong' value of the <INPUT> field instead of the 'valid' value of the <input> field.
> How I achieved this situation:
> Even though MyFaces typically renders input fields with lower case characters (XHTML), when a page is rendered by Firefox, the nodeName of input elements seems to be the upper case variant. I wonder why.
> Now, we use RichFaces' a4j:poll in order to perform periodical partial page updates.
> The updated part of the page may contain locically equivalent input fields, but in lower case.
> The JavaScript behind a4j:poll uses the W3C DOM method replaceChild() in order to update page fragments.
> This method replaces the original <INPUT> fields by <input> fields.
> Of course, it doesn't matter how <input> elements become part of a page.
> In oamSetHiddenInput() <input> should be treated like <INPUT>.

--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

[My Faces - Dev mailing list]

     [ https://issues.apache.org/jira/browse/MYFACES-2256?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Leonardo Uribe resolved MYFACES-2256.
-------------------------------------

       Resolution: Fixed
    Fix Version/s: 1.2.8-SNAPSHOT
         Assignee: Leonardo Uribe

Thanks to Lutz Ulruch for this patch

> oamSetHiddenInput should not ignore nodeName=='input' (lower case characters)
> -----------------------------------------------------------------------------
>
>                 Key: MYFACES-2256
>                 URL: https://issues.apache.org/jira/browse/MYFACES-2256
>             Project: MyFaces Core
>          Issue Type: Bug
>    Affects Versions: 1.2.6
>         Environment: Windows XP Professional, Firefox 3.0.11, MyFaces 1.2.6 Tomahawk-12 1.1.8 (and RichFaces 3.3.1 GA), Tomcat 6.
>            Reporter: Lutz Ulruch
>            Assignee: Leonardo Uribe
>             Fix For: 1.2.8-SNAPSHOT
>
>
> In
> function oamSetHiddenInput(formname, name, value)
> the test
> if(typeof form.elements[name]!='undefined' && form.elements[name].nodeName=='INPUT')
> does not cover <input> elements where nodeName is 'input' (that is: lower case characters instead of upper case 'INPUT').
> This results in the following error:
> If the form already contains an input element with the specified name, but nodeName=='input', the test fails and a new input element with the same name is added to the form.
> If the form is submitted, the HTTP request will contain 2 values for the parameter referenced by 'name':
> one is the value of the unchanged <INPUT> field, the other is the value of the newly created <input> field.
> Most UIComponents do not check if there is more than one value for a parameter.
> Probably they will just interpret the first value - which might be the 'wrong' value of the <INPUT> field instead of the 'valid' value of the <input> field.
> How I achieved this situation:
> Even though MyFaces typically renders input fields with lower case characters (XHTML), when a page is rendered by Firefox, the nodeName of input elements seems to be the upper case variant. I wonder why.
> Now, we use RichFaces' a4j:poll in order to perform periodical partial page updates.
> The updated part of the page may contain locically equivalent input fields, but in lower case.
> The JavaScript behind a4j:poll uses the W3C DOM method replaceChild() in order to update page fragments.
> This method replaces the original <INPUT> fields by <input> fields.
> Of course, it doesn't matter how <input> elements become part of a page.
> In oamSetHiddenInput() <input> should be treated like <INPUT>.

--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.