|Syntactic Confectionery Delight|
Re: Undocumented Operator Japhby chargrill (Parson)
|on Nov 21, 2006 at 22:52 UTC||Need Help??|
Here's a crack at deobfuscating it:
First things first, fix that pesky whitespace.
Clean up a few things, simplifying where possible:
But there's still those extra "+"'s and "-"'s in that second line... Well, thanks to that hint about blazar's post, I think I know what's going on, so let me try something...
So here's what's going on: line 1 sets the @= array to some values, which happen to be the equivalent of "( 'Gqw', 'pWr', 'lpl', 'bnT', 'bnp', 'Dwg', 'hav' )".
Line 2 cleverly "re-maps" (using a map inside a map, nonetheless) the @= array by ++ "incrementing" every value 2,127 times - the exact number of times required to "increment" Gqw to Jus*. The preceeding @+ - just adds a little line noise, as it doesn't seem to affect the final result if removed. Same with the + in front of $= (which, incidentally, (as documented ;) defaults to 60).
Line 3 joins @= by the value of $*, which is (as documented ;) undefined, giving us "JustAnotherPerlHacker", which is then split in between a lowercase character and a following uppercase character thanks to the zero width positive look-ahead assertion /(?=[A-Q])/, and the resulting list from the split is stored in @'.
Finally, with a few + cluttering things up (and used as qq's quoting character for interpolating the values of the variables in the string given to die), you 'print' the contents of @', trusting that (as documented ;) the default value of a single space for $" hasn't changed, and $/ also defaults (as documented ;) to a newline.
* (The bit of perldoc perlop that details this is: The auto-increment operator has a little extra builtin magic to it. ... If, however, the variable has been used in only string contexts since it was set, and has a value that is not the empty string and matches the pattern "/^[a-zA-Z]*[0-9]*\z/", the increment is done as a string, preserving each character within its range, with carry .... Incrementing "Gqw", and you get "Gqx")
Update: Cleared up the "obfuscated deobfuscation" as noted by ysth, though the original test was written to most closely resemble (from at least my point of view) the original obfuscated code. Apologies for any confusion. Also added a little extra explanation (with reference to associated documentation).