Dale Newfield wrote:
Martin Gainty wrote:
this is the applicable code in my copy of optiontransferselect.ftl
Thanks for making me look at this--it led me straight to my problem!
So we know there is a problem with the freemarker libraries, and that
it's either in ?exists, or stack.findString(), or in the assumption that
the first being true implies the second will always produce a valid
result. Is it true that if it has a value that value must be
extractable as a string? The freemarker docs suggest that maybe
?has_content would be better than ?exists ...
This rational is almost right, but has a big hole. We're asking whether
the tag has a parameter set, and then assuming the data has not null for
that attribute. I indeed had a piece of data with null for that value,
which caused my problem. Since that attribute was derived, I was able
to modify my model object to return "" in that case instead of null, and
I'm golden. I might suggest this alteration to the .ftl file, though:
<#if parameters.doubleListValue?exists><#t/>
+ <#if
stack.findString(parameters.doubleListValue)?has_content><#t/>
<#assign doubleItemValue =
stack.findString(parameters.doubleListValue) /><#t/>
+ </#if><#t/>
<#else><#t/>
<#assign doubleItemValue = stack.findString('top') /><#t/>
</#if><#t/>
Now I don't have to change interfaces for large groups! Yay! Thanks!
-Dale
P.S.: I shoulda known that 1124 was not a magic enough number to be a
systemic problem--that number should have led me to look at my data
sooner :-/.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]