|There's more than one way to do things|
RE: RE: Tricks with tr///by btrott (Parson)
|on Feb 23, 2000 at 01:04 UTC||Need Help??|
Well, the comments describe what the code actually does, but I guess you're asking *why* it does what it does.
Let's take one particular statement and go over it--since they all have a similar structure, I think that'll help you to figure out the rest.
Let's take this one:
The comments provided say that this "squishes" a string. If you actually try this out in a little test program, you'll see exactly what it does:
This prints "bokeper"--in other words, it gets rid of consecutive letters ("oo" => "o").
Why does it do that? Let's write the statement in a slightly different way:
This is the same statement, because:
"y" is a synonym for "tr";
and the "tr" operator can use any character to separate the SEARCHLIST from the REPLACEMENTLIST.
So in the obfuscated example, "tr" is "y", and the separator is "s". We can replace those to make it look more "normal", just as we did above.
Why all the "c"'s at the end, though? Well, we can take most of those off, too:
And now we no longer have obfuscated code: this is just the regular "tr" operator, working as advertised.
The "c" tells "tr" to complement the SEARCHLIST; the SEARCHLIST is an empty string, so the complement is any character--so we're searching for any character. The "s" tells "tr" to "squish" the string-- from perlman:perlop:
In other words, "oo" => "o", and so on.
The final thing going on is that if the REPLACEMENTLIST is empty (as it is here), the SEARCHLIST is replicated (so the SEARCHLIST and REPLACEMENTLIST are the same).
So we're translating every character (the complement of an empty SEARCHLIST) into itself, *unless* there's a sequence of characters that translate to the same character, in which case we "squish" the sequence down to one instance of that character.
So, I hope this helps. Try to figure out the rest yourself, using some test programs and perlman:perlop.
And if I've made any mistakes, someone please correct them. :)