Writing big XML files where beginning depends on end.

[Magnus Lycka]

Fredrik Lundh wrote:
> what's the typical overall structure for this tree ?  is it short and wide, or tall and
> narrow ?

That varies quite a bit. It's basically a call graph for a
functional language, where also aggregated results are stored
in the XML tree. In other words, the shape depends on the
program we're analyzing. With a loop in the top level, we
could get plenty of nodes in the root, but we need to handle
different cases. Typically, the language in question isn't
used to build very complex things, so this hasn't really
been an issue before...

I tried to look at a 100MB xml file in a browser, to see the
shape, but Mozilla died after growing to several GB in size...