Rene de Visser wrote:
If I remember correctly, the data type of the tree in HXT is something like
data Tree = Tree NodeData [Tree]
which means that already processed parts of the tree can't be garbage collected because the parent node is holding onto them.
This statement only holds, if you still need the root node after processing. When the tree is processed top down, the root and all trees to the left of the currently processed tree can be garbage collected. This is precisely the type of computation needed when e.g. working with picklers to convert XML into native Haskell data structures. BTW: I've taken the tagsoup lib and wrote a small parser to build a tree out of the stream of tags. It's about a 100 lines of code. This DOM parser does not need to read until the closing tag to build an element node, so it should be as lasy as possible. A first version for HTML already runs on my box, but it stil needs a bit of testing Uwe