tapestry-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mindbridge" <mindbridge...@yahoo.com>
Subject Re: Table component changes checked in
Date Wed, 07 Jan 2004 01:51:40 GMT
Hi,

Just some final observations:

> > - Shouldn't the 'type' attribute of <parameter> in the
> > component specs be required? Or perhaps have it default to
> > java.lang.Object?
>
> That's how I thought it worked. It doesn't?

Yes, it seems it does. Not sure what I did earlier, must have been
confusion.

> > - Shouldn't passing a parameter to a component when it is
> > marked as 'reserved' cause an exception? At the moment it
> > does not (thank god for Spindle!).
>
> The correct behavior is to ignore static bindings that originate in the
HTML template, but throw an
> exception for non-static bindings, or static bindings in the
specification.

I did some checking and this is almost correct.
If the parameter value is defined in the template:
  - an exception is thrown if the value is an OGNL expression or a message
  - no exception is thrown if the value is static (a normal string)
This is as expected, I guess.

But if the parameter value is defined in the spec, no exception is thrown in
all cases.
See PageLoader.java line 367.

Is this an omission?

> Any non-null string should evaluate to true.  The problem is that 'false'
also evaluates to true,
> since it is also a non-null string. This is a very, very common error ...
perhaps we should special
> case 'false' to evaluate to true (inside Tapestry.evaluateBoolean() ) ?

Done, change has been checked in.

Best regards,
-mb


---------------------------------------------------------------------
To unsubscribe, e-mail: tapestry-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tapestry-dev-help@jakarta.apache.org


Mime
View raw message