August 12th, 2010

Serialization Fixed Up

I finally solved yesterday's serialization problem. The trick was that I'm doing several things that are really oddball in order to cause the old serialized templates to create new different serialized templates as they are serialized back in. And the schema numbers weren't quite compatible between some of the different derived types that I used to have and am now busily eliminating.

But since I can figure out what the type of data was that was being serialized in the template and whether I was serializing a scalar or an array value by testing a static function I popped into the wrapper class that the templates are serializing, I can pick and choose my schema ranges depending on what the original class was that I was serializing.

So now it works.