fix for negative weights from Hto4l - changing all weights by *-1

Hto4l takes an incoming LHE PowHeg event with a Higgs, and merges in the H4l decay. In the original implementation, the mode we (H4l) were using has NLOEW and multi-photon parton shower and there are almost no negative weights. So I didn't include treatment for negative weights as I thought that there were none. When running at test for just NLOEW, which has several % of negative weights, I realized that the negative weights were not propagated to the weight block. So I have now implemented this. And I also set the weight in the event block (XWGTUP) to be the weight for id=0. I had seen a comment that pythia may use XWGTUP, but I checked this and didn't see any effect if XWGTUP is +/-1 or the same as the PowHeg nominal x-sec in id=0. So this shouldn't have any effect. But setting XWGTUP to be the id=0 weight seems more 'correct'. This is how the Prophecy4f implementation was done.