Validator validwhen cannot test between two different indices in array

• Key: STR-2802
• Type: Bug
• Status: Closed
• Resolution: Fixed
• Priority: Major
• Assignee: Unassigned
• Reporter: Oudot Christophe
• Component/s: Core
• Affects Version/s: 1.2.8
• Fix Version/s: 1.3.10, 1.4.0

Environment

Operating System: Windows 2000
Platform: PC

Description

This code in the validator.xml fails with this error : "unexpected token adr.."

  
<field property="adr[0]" depends ="validwhen">
  <var>
      <var-name>test</var-name>
      <var-value>((*this*!=null) or (adr[1] != null ))</var-value>
  </var>
</field>


adr is an array of two adress lines. Only one of them needs tobe filled.
According to the document of the validator, it should work.

Comments

[Niall Pemberton]

Adding a link to the related thread on the user list:

http://www.mail-archive.com/user%40struts.apache.org/msg43177.html

I had a quick look at the source (ValidWhenParser.g) and I agree it looks like
it should cater for this. I also tried adding a test case (to
TestValidWhen.java) to and got the same result as Christophe.

[Eric Rizzo]

Is there any workaround for this bug? I have a similar requirement (need to check for presence of EITHER one field or at least one item in an array field (items selected from an HTML list), but have run into the same issue.
Is there any other way to configure the validator to do this kind of check, at least until this bug gets fixed?

[Laird Nelson]

It's now March 2007; I was wondering if any progress had been made on this bug. I am working on struts 1.3.5, and I encounter the same issue with the following code:

<var-value>((*this* != null) or ((govIDParts[0] == null) and (govIDParts[2] == null)))</var-value>

[Niall Pemberton]

I doubt anyones looking at this - It needs someone to come forward with a patch to fix it

[Henri Yandell]

Do you still have your test case Niall?

[Niall Pemberton]

I didn't (I changed machines since then) - but I've re-created one and committed it:

http://svn.apache.org/viewvc?view=rev&revision=515237

I suspect that the third of the array rules shouldn't have the LBRACKET on the end of it. The problem with both Niall's test and the original report is that neither of them match one of the 5 array rules.

ie:

Index: ValidWhenParser.g
===================================================================
--- ValidWhenParser.g (revision 518754)
+++ ValidWhenParser.g (working copy)
@@ -182,7 +182,7 @@
             Object i3 = argStack.pop();
             argStack.push(ValidatorUtils.getValueAsString(form, i3 + "[" + i4 + "]" + i5));
 }
-| identifier LBRACKET integer RBRACKET LBRACKET {
+| identifier LBRACKET integer RBRACKET {
             Object i7 = argStack.pop();
             Object i6 = argStack.pop();
             argStack.push(ValidatorUtils.getValueAsString(form, i6 + "[" + i7 + "]"));

[Henri Yandell]

 svn ci -m "Fixing STR-2802 by removing the unnecessary LBRACKET. Test now passes and I don't understan
d what the LBRACKET would have been there for so I think it was just an error that's been in there forever and ever. " src/
Sending src/main/java/org/apache/struts/validator/validwhen/ValidWhenParser.g
Sending src/main/java/org/apache/struts/validator/validwhen/ValidWhenParser.java
Transmitting file data ..
Committed revision 520763.

[Niall Pemberton]

The fix for STR-3088 has also back-ported the fix for this issue into the 1.3 Branch - see revision 569606:

http://svn.apache.org/viewvc?view=rev&revision=569606