On 13-10-28 1:59 PM, Mick Semb Wever wrote:
In the meantime, I wrote up a hacky Aspect to intercept the
RuntimeException and throw the cause instead when OptionRenderer is
the caller.
I'll be interested in anything you think makes life easier to debug such
problems. I know OptionsRenderer can too easily mask the underlying
problems.
Honestly, not entirely sure. Took me a while to step through everything to get a handle on what was happening. Too many contexts, factories,
initializers, etc for the un-initiated to really follow easily. All my Aspect does at the moment is intercept the SpringWildercardContext
.getResources() method and throws the root cause instead of the RuntimeException if it is being called by OptionsRenderer. But this type of hack is
even worse to actually debug, so would not begin to even suggest something like that.
Can you please update your Blog with the relevant information in the
meantime? ie: ensuring the "cascade=true" is present in the
list-attributes and indicate that it requires 3.0.2-SNAPSHOT at the
minimum to be functional?
Done for cascade=true.
The article afaik doesn't refer anywhere to specific tiles-3 artifacts,
so i'd like to leave it to the reader that they should use the latest
available and get tiles-3.0.2 out quickly. (I have to get tiles-master-6
out first).
I don't actually see a tiles-3.0.2-SNAPSHOT anywhere. The pom version for
extras is 3.1-SNAPSHOT. Is there even going to be a 3.0.2 relelase?
I think that minimum version numbers in your blog could be very helpful. So that people at least know/understand that to use the built-in classes,
they need at least Spring 3.2 and Tiles 3.0.2. Otherwise, they will be spending hours (ie: me! :)) pulling hair trying to understand why things
aren't working.
