It's not really the just the source he's showing though, it's the source of the subexpressions leading to the failure of the last expression.
Working on the string level is even more confusing in this case, it's beyond just parsing perl without losing sanity, you must also reconstruct it. I have no idea if this rewriting of functions is at all possible, but here's an idea: I would think the best way (in theory) to go about this is to deconstruct the last expression of the closure, and reconstruct it such that the values of every subexpression is cached in a variable, and the variable replaces the subexpression. The variables live outside the new closure, which closes over them. The subexpressions are also deparsed individually. Then on failure the subexpression variable hash thingy could just be dumped. Thoughts? -- Yuval Kogman <[EMAIL PROTECTED]> http://nothingmuch.woobling.org 0xEBD27418
