hey, well, I've taken a look at the refactored scandeps code, and yes, there is one discrepancy.
As I conceived it, -c did runtime analysis of the code, -e furthered that by actually running the code and analyzing what came out of that. However, it was conceived as an *add on* to the static scanning that happens by default. (in fact, I had a flag - -n - which explicitly turned *off* static scanning). As it stands however, its an either/or thing: either you get static scanning or you get runtime scanning. IMO it should do this instead. -c is a good 'catch-all' which picks up things that static scanning can miss, but if you have a single require anywhere in your code, you'll miss it without static scanning. Its benefits are that it produces small, small executables, and handles weird cases, but I don't want to be forced to forego static scanning just because I need -c. In fact, I think there should be a third type of scanning, namely run-time static. Where the *only* thing it picks up are things like 'require ...'. This way, you could have small executable sizes and not miss requires, and you wouldn't have to actually run the script, ether. Ed
