Re: generic grouping without nesting for-each-group?

On Wed, Jul 28, 2010 at 15:29, Martin Honnen <Martin.Honnen@xxxxxx> wrote:
> It has become rather complicated with two functions due to the requirement
> to add missing levels.

It hadn't actually occurred to me to do the recursion in a function
rather than a named template. Fun!

> Your latest suggestion to first normalize the input by adding missing levels
> and then to group in a second pass might indeed be an approach leading to
> shorter and less complicated code.

So far not thought of a better suggestion... though I have a natural
suspicion of adding in phantom structure in any case.

-James